Rebind textures if format changes or they're buffer textures

* Allocate work buffer for audio renderer instead of using guest supplied memory

* Typo

* Use GC.AllocateArray to allocate pinned array

.github/ISSUE_TEMPLATE/bug_report.md

@@ -1,6 +1,6 @@
 ---
 name: Bug Report
-about: Something doesn't work correctly in Ryujinx.
+about: Something doesn't work correctly in Ryujinx. Note that game-specific issues should be instead posted on the Game Compatibility List at https://github.com/Ryujinx/Ryujinx-Games-List, unless it is a provable regression.
 #assignees:
 ---
 

ARMeilleure/CodeGen/Linking/Symbol.cs

@@ -58,7 +58,6 @@ namespace ARMeilleure.CodeGen.Linking
         /// <param name="a">First instance</param>
         /// <param name="b">Second instance</param>
         /// <returns><see langword="true"/> if not equal; otherwise <see langword="false"/></returns>
-        /// <inheritdoc/>
         public static bool operator !=(Symbol a, Symbol b)
         {
             return !(a == b);

ARMeilleure/CodeGen/RegisterAllocators/StackAllocator.cs

@@ -1,6 +1,4 @@
-using ARMeilleure.Common;
 using ARMeilleure.IntermediateRepresentation;
-using System;
 
 namespace ARMeilleure.CodeGen.RegisterAllocators
 {

ARMeilleure/CodeGen/X86/AssemblerTable.cs

@@ -4,6 +4,11 @@ namespace ARMeilleure.CodeGen.X86
 {
     partial class Assembler
     {
+        public static bool SupportsVexPrefix(X86Instruction inst)
+        {
+            return _instTable[(int)inst].Flags.HasFlag(InstructionFlags.Vex);
+        }
+
         private const int BadOp = 0;
 
         [Flags]
@@ -152,6 +157,7 @@ namespace ARMeilleure.CodeGen.X86
             Add(X86Instruction.Paddd,        new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x00000ffe, InstructionFlags.Vex | InstructionFlags.Prefix66));
             Add(X86Instruction.Paddq,        new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x00000fd4, InstructionFlags.Vex | InstructionFlags.Prefix66));
             Add(X86Instruction.Paddw,        new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x00000ffd, InstructionFlags.Vex | InstructionFlags.Prefix66));
+            Add(X86Instruction.Palignr,      new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x000f3a0f, InstructionFlags.Vex | InstructionFlags.Prefix66));
             Add(X86Instruction.Pand,         new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x00000fdb, InstructionFlags.Vex | InstructionFlags.Prefix66));
             Add(X86Instruction.Pandn,        new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x00000fdf, InstructionFlags.Vex | InstructionFlags.Prefix66));
             Add(X86Instruction.Pavgb,        new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x00000fe0, InstructionFlags.Vex | InstructionFlags.Prefix66));
@@ -234,6 +240,9 @@ namespace ARMeilleure.CodeGen.X86
             Add(X86Instruction.Rsqrtss,      new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x00000f52, InstructionFlags.Vex | InstructionFlags.PrefixF3));
             Add(X86Instruction.Sar,          new InstructionInfo(0x070000d3, 0x070000c1, BadOp,      BadOp,      BadOp,      InstructionFlags.None));
             Add(X86Instruction.Setcc,        new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x00000f90, InstructionFlags.Reg8Dest));
+            Add(X86Instruction.Sha256Msg1,   new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x000f38cc, InstructionFlags.None));
+            Add(X86Instruction.Sha256Msg2,   new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x000f38cd, InstructionFlags.None));
+            Add(X86Instruction.Sha256Rnds2,  new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x000f38cb, InstructionFlags.None));
             Add(X86Instruction.Shl,          new InstructionInfo(0x040000d3, 0x040000c1, BadOp,      BadOp,      BadOp,      InstructionFlags.None));
             Add(X86Instruction.Shr,          new InstructionInfo(0x050000d3, 0x050000c1, BadOp,      BadOp,      BadOp,      InstructionFlags.None));
             Add(X86Instruction.Shufpd,       new InstructionInfo(BadOp,      BadOp,      BadOp,      BadOp,      0x00000fc6, InstructionFlags.Vex | InstructionFlags.Prefix66));

ARMeilleure/CodeGen/X86/CallingConvention.cs

@@ -1,5 +1,4 @@
 using System;
-using System.Runtime.InteropServices;
 
 namespace ARMeilleure.CodeGen.X86
 {

ARMeilleure/CodeGen/X86/HardwareCapabilities.cs

@@ -12,21 +12,28 @@ namespace ARMeilleure.CodeGen.X86
                 return;
             }
 
-            (_, _, int ecx, int edx) = X86Base.CpuId(0x00000001, 0x00000000);
+            (int maxNum, _, _, _) = X86Base.CpuId(0x00000000, 0x00000000);
 
-            FeatureInfoEdx = (FeatureFlagsEdx)edx;
-            FeatureInfoEcx = (FeatureFlagsEcx)ecx;
+            (_, _, int ecx1, int edx1) = X86Base.CpuId(0x00000001, 0x00000000);
+            FeatureInfo1Edx = (FeatureFlags1Edx)edx1;
+            FeatureInfo1Ecx = (FeatureFlags1Ecx)ecx1;
+
+            if (maxNum >= 7)
+            {
+                (_, int ebx7, _, _) = X86Base.CpuId(0x00000007, 0x00000000);
+                FeatureInfo7Ebx = (FeatureFlags7Ebx)ebx7;
+            }
         }
 
         [Flags]
-        public enum FeatureFlagsEdx
+        public enum FeatureFlags1Edx
         {
             Sse = 1 << 25,
             Sse2 = 1 << 26
         }
 
         [Flags]
-        public enum FeatureFlagsEcx
+        public enum FeatureFlags1Ecx
         {
             Sse3 = 1 << 0,
             Pclmulqdq = 1 << 1,
@@ -40,21 +47,31 @@ namespace ARMeilleure.CodeGen.X86
             F16c = 1 << 29
         }
 
-        public static FeatureFlagsEdx FeatureInfoEdx { get; }
-        public static FeatureFlagsEcx FeatureInfoEcx { get; }
+        [Flags]
+        public enum FeatureFlags7Ebx
+        {
+            Avx2 = 1 << 5,
+            Sha = 1 << 29
+        }
 
-        public static bool SupportsSse => FeatureInfoEdx.HasFlag(FeatureFlagsEdx.Sse);
-        public static bool SupportsSse2 => FeatureInfoEdx.HasFlag(FeatureFlagsEdx.Sse2);
-        public static bool SupportsSse3 => FeatureInfoEcx.HasFlag(FeatureFlagsEcx.Sse3);
-        public static bool SupportsPclmulqdq => FeatureInfoEcx.HasFlag(FeatureFlagsEcx.Pclmulqdq);
-        public static bool SupportsSsse3 => FeatureInfoEcx.HasFlag(FeatureFlagsEcx.Ssse3);
-        public static bool SupportsFma => FeatureInfoEcx.HasFlag(FeatureFlagsEcx.Fma);
-        public static bool SupportsSse41 => FeatureInfoEcx.HasFlag(FeatureFlagsEcx.Sse41);
-        public static bool SupportsSse42 => FeatureInfoEcx.HasFlag(FeatureFlagsEcx.Sse42);
-        public static bool SupportsPopcnt => FeatureInfoEcx.HasFlag(FeatureFlagsEcx.Popcnt);
-        public static bool SupportsAesni => FeatureInfoEcx.HasFlag(FeatureFlagsEcx.Aes);
-        public static bool SupportsAvx => FeatureInfoEcx.HasFlag(FeatureFlagsEcx.Avx);
-        public static bool SupportsF16c => FeatureInfoEcx.HasFlag(FeatureFlagsEcx.F16c);
+        public static FeatureFlags1Edx FeatureInfo1Edx { get; }
+        public static FeatureFlags1Ecx FeatureInfo1Ecx { get; }
+        public static FeatureFlags7Ebx FeatureInfo7Ebx { get; } = 0;
+
+        public static bool SupportsSse => FeatureInfo1Edx.HasFlag(FeatureFlags1Edx.Sse);
+        public static bool SupportsSse2 => FeatureInfo1Edx.HasFlag(FeatureFlags1Edx.Sse2);
+        public static bool SupportsSse3 => FeatureInfo1Ecx.HasFlag(FeatureFlags1Ecx.Sse3);
+        public static bool SupportsPclmulqdq => FeatureInfo1Ecx.HasFlag(FeatureFlags1Ecx.Pclmulqdq);
+        public static bool SupportsSsse3 => FeatureInfo1Ecx.HasFlag(FeatureFlags1Ecx.Ssse3);
+        public static bool SupportsFma => FeatureInfo1Ecx.HasFlag(FeatureFlags1Ecx.Fma);
+        public static bool SupportsSse41 => FeatureInfo1Ecx.HasFlag(FeatureFlags1Ecx.Sse41);
+        public static bool SupportsSse42 => FeatureInfo1Ecx.HasFlag(FeatureFlags1Ecx.Sse42);
+        public static bool SupportsPopcnt => FeatureInfo1Ecx.HasFlag(FeatureFlags1Ecx.Popcnt);
+        public static bool SupportsAesni => FeatureInfo1Ecx.HasFlag(FeatureFlags1Ecx.Aes);
+        public static bool SupportsAvx => FeatureInfo1Ecx.HasFlag(FeatureFlags1Ecx.Avx);
+        public static bool SupportsAvx2 => FeatureInfo7Ebx.HasFlag(FeatureFlags7Ebx.Avx2) && SupportsAvx;
+        public static bool SupportsF16c => FeatureInfo1Ecx.HasFlag(FeatureFlags1Ecx.F16c);
+        public static bool SupportsSha => FeatureInfo7Ebx.HasFlag(FeatureFlags7Ebx.Sha);
 
         public static bool ForceLegacySse { get; set; }
 

ARMeilleure/CodeGen/X86/IntrinsicTable.cs

@@ -82,6 +82,7 @@ namespace ARMeilleure.CodeGen.X86
             Add(Intrinsic.X86Paddd,        new IntrinsicInfo(X86Instruction.Paddd,        IntrinsicType.Binary));
             Add(Intrinsic.X86Paddq,        new IntrinsicInfo(X86Instruction.Paddq,        IntrinsicType.Binary));
             Add(Intrinsic.X86Paddw,        new IntrinsicInfo(X86Instruction.Paddw,        IntrinsicType.Binary));
+            Add(Intrinsic.X86Palignr,      new IntrinsicInfo(X86Instruction.Palignr,      IntrinsicType.TernaryImm));
             Add(Intrinsic.X86Pand,         new IntrinsicInfo(X86Instruction.Pand,         IntrinsicType.Binary));
             Add(Intrinsic.X86Pandn,        new IntrinsicInfo(X86Instruction.Pandn,        IntrinsicType.Binary));
             Add(Intrinsic.X86Pavgb,        new IntrinsicInfo(X86Instruction.Pavgb,        IntrinsicType.Binary));
@@ -151,6 +152,9 @@ namespace ARMeilleure.CodeGen.X86
             Add(Intrinsic.X86Roundss,      new IntrinsicInfo(X86Instruction.Roundss,      IntrinsicType.BinaryImm));
             Add(Intrinsic.X86Rsqrtps,      new IntrinsicInfo(X86Instruction.Rsqrtps,      IntrinsicType.Unary));
             Add(Intrinsic.X86Rsqrtss,      new IntrinsicInfo(X86Instruction.Rsqrtss,      IntrinsicType.Unary));
+            Add(Intrinsic.X86Sha256Msg1,   new IntrinsicInfo(X86Instruction.Sha256Msg1,   IntrinsicType.Binary));
+            Add(Intrinsic.X86Sha256Msg2,   new IntrinsicInfo(X86Instruction.Sha256Msg2,   IntrinsicType.Binary));
+            Add(Intrinsic.X86Sha256Rnds2,  new IntrinsicInfo(X86Instruction.Sha256Rnds2,  IntrinsicType.Ternary));
             Add(Intrinsic.X86Shufpd,       new IntrinsicInfo(X86Instruction.Shufpd,       IntrinsicType.TernaryImm));
             Add(Intrinsic.X86Shufps,       new IntrinsicInfo(X86Instruction.Shufps,       IntrinsicType.TernaryImm));
             Add(Intrinsic.X86Sqrtpd,       new IntrinsicInfo(X86Instruction.Sqrtpd,       IntrinsicType.Unary));

ARMeilleure/CodeGen/X86/PreAllocator.cs

@@ -308,11 +308,13 @@ namespace ARMeilleure.CodeGen.X86
 
                 case Instruction.Extended:
                 {
-                    // BLENDVPD, BLENDVPS, PBLENDVB last operand is always implied to be XMM0 when VEX is not supported.
-                    if ((node.Intrinsic == Intrinsic.X86Blendvpd ||
+                    bool isBlend = node.Intrinsic == Intrinsic.X86Blendvpd ||
                                    node.Intrinsic == Intrinsic.X86Blendvps ||
-                         node.Intrinsic == Intrinsic.X86Pblendvb) &&
-                         !HardwareCapabilities.SupportsVexEncoding)
+                                   node.Intrinsic == Intrinsic.X86Pblendvb;
+
+                    // BLENDVPD, BLENDVPS, PBLENDVB last operand is always implied to be XMM0 when VEX is not supported.
+                    // SHA256RNDS2 always has an implied XMM0 as a last operand.
+                    if ((isBlend && !HardwareCapabilities.SupportsVexEncoding) || node.Intrinsic == Intrinsic.X86Sha256Rnds2)
                     {
                         Operand xmm0 = Xmm(X86Register.Xmm0, OperandType.V128);
 
@@ -1297,11 +1299,15 @@ namespace ARMeilleure.CodeGen.X86
         {
             if (IsIntrinsic(operation.Instruction))
             {
+                IntrinsicInfo info = IntrinsicTable.GetInfo(operation.Intrinsic);
+
+                bool hasVex = HardwareCapabilities.SupportsVexEncoding && Assembler.SupportsVexPrefix(info.Inst);
+
                 bool isUnary = operation.SourcesCount < 2;
 
                 bool hasVecDest = operation.Destination != default && operation.Destination.Type == OperandType.V128;
 
-                return !HardwareCapabilities.SupportsVexEncoding && !isUnary && hasVecDest;
+                return !hasVex && !isUnary && hasVecDest;
             }
 
             return false;

ARMeilleure/CodeGen/X86/X86Instruction.cs

@@ -98,6 +98,7 @@ namespace ARMeilleure.CodeGen.X86
         Paddd,
         Paddq,
         Paddw,
+        Palignr,
         Pand,
         Pandn,
         Pavgb,
@@ -180,6 +181,9 @@ namespace ARMeilleure.CodeGen.X86
         Rsqrtss,
         Sar,
         Setcc,
+        Sha256Msg1,
+        Sha256Msg2,
+        Sha256Rnds2,
         Shl,
         Shr,
         Shufpd,

ARMeilleure/Common/AddressTable.cs

@@ -1,4 +1,4 @@
-using ARMeilleure.Diagnostics.EventSources;
+using ARMeilleure.Diagnostics;
 using System;
 using System.Collections.Generic;
 using System.Runtime.InteropServices;
@@ -206,7 +206,7 @@ namespace ARMeilleure.Common
         /// <typeparam name="T">Type of elements</typeparam>
         /// <param name="length">Number of elements</param>
         /// <param name="fill">Fill value</param>
-        /// <param name="leaf"><see langword="true"/> if leaf; otherwise <see langword=""="false"/></param>
+        /// <param name="leaf"><see langword="true"/> if leaf; otherwise <see langword="false"/></param>
         /// <returns>Allocated block</returns>
         private IntPtr Allocate<T>(int length, T fill, bool leaf) where T : unmanaged
         {
@@ -218,7 +218,7 @@ namespace ARMeilleure.Common
 
             _pages.Add(page);
 
-            AddressTableEventSource.Log.Allocated(size, leaf);
+            TranslatorEventSource.Log.AddressTableAllocated(size, leaf);
 
             return page;
         }

ARMeilleure/Common/ArenaAllocator.cs

@@ -1,7 +1,6 @@
 using System;
 using System.Collections.Generic;
 using System.Runtime.CompilerServices;
-using System.Threading;
 
 namespace ARMeilleure.Common
 {

ARMeilleure/Common/Counter.cs

@@ -9,6 +9,9 @@ namespace ARMeilleure.Common
     class Counter<T> : IDisposable where T : unmanaged
     {
         private bool _disposed;
+        /// <summary>
+        /// Index in the <see cref="EntryTable{T}"/>
+        /// </summary>
         private readonly int _index;
         private readonly EntryTable<T> _countTable;
 
@@ -17,7 +20,6 @@ namespace ARMeilleure.Common
         /// <see cref="EntryTable{T}"/> instance and index.
         /// </summary>
         /// <param name="countTable"><see cref="EntryTable{T}"/> instance</param>
-        /// <param name="index">Index in the <see cref="EntryTable{T}"/></param>
         /// <exception cref="ArgumentNullException"><paramref name="countTable"/> is <see langword="null"/></exception>
         /// <exception cref="ArgumentException"><typeparamref name="T"/> is unsupported</exception>
         public Counter(EntryTable<T> countTable)
@@ -68,7 +70,7 @@ namespace ARMeilleure.Common
         /// <summary>
         /// Releases all unmanaged and optionally managed resources used by the <see cref="Counter{T}"/> instance.
         /// </summary>
-        /// <param name="disposing"><see langword="true"/> to dispose managed resources also; otherwise just unmanaged resouces</param>
+        /// <param name="disposing"><see langword="true"/> to dispose managed resources also; otherwise just unmanaged resources</param>
         protected virtual void Dispose(bool disposing)
         {
             if (!_disposed)

ARMeilleure/Decoders/Decoder.cs

@@ -251,6 +251,13 @@ namespace ARMeilleure.Decoders
                 return false;
             }
 
+            // Compare and branch instructions are always conditional.
+            if (opCode.Instruction.Name == InstName.Cbz ||
+                opCode.Instruction.Name == InstName.Cbnz)
+            {
+                return false;
+            }
+
             // Note: On ARM32, most instructions have conditional execution,
             // so there's no "Always" (unconditional) branch like on ARM64.
             // We need to check if the condition is "Always" instead.

ARMeilleure/Decoders/IOpCode32Mem.cs

@@ -3,6 +3,7 @@ namespace ARMeilleure.Decoders
     interface IOpCode32Mem : IOpCode32
     {
         int Rt { get; }
+        int Rt2 => Rt | 1;
         int Rn { get; }
 
         bool WBack { get; }

ARMeilleure/Decoders/IOpCode32MemRsImm.cs

@@ -0,0 +1,8 @@
+namespace ARMeilleure.Decoders
+{
+    interface IOpCode32MemRsImm : IOpCode32Mem
+    {
+        int Rm { get; }
+        ShiftType ShiftType { get; }
+    }
+}

ARMeilleure/Decoders/OpCode32MemMult.cs

@@ -1,3 +1,5 @@
+using System.Numerics;
+
 namespace ARMeilleure.Decoders
 {
     class OpCode32MemMult : OpCode32, IOpCode32MemMult
@@ -23,14 +25,7 @@ namespace ARMeilleure.Decoders
 
             RegisterMask = opCode & 0xffff;
 
-            int regsSize = 0;
-
-            for (int index = 0; index < 16; index++)
-            {
-                regsSize += (RegisterMask >> index) & 1;
-            }
-
-            regsSize *= 4;
+            int regsSize = BitOperations.PopCount((uint)RegisterMask) * 4;
 
             if (!u)
             {

ARMeilleure/Decoders/OpCode32MemRsImm.cs

@@ -1,6 +1,6 @@
 namespace ARMeilleure.Decoders
 {
-    class OpCode32MemRsImm : OpCode32Mem
+    class OpCode32MemRsImm : OpCode32Mem, IOpCode32MemRsImm
     {
         public int Rm { get; }
         public ShiftType ShiftType { get; }

ARMeilleure/Decoders/OpCode32Mrs.cs

@@ -0,0 +1,16 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCode32Mrs : OpCode32
+    {
+        public bool R  { get; }
+        public int  Rd { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCode32Mrs(inst, address, opCode);
+
+        public OpCode32Mrs(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            R = ((opCode >> 22) & 1) != 0;
+            Rd = (opCode >> 12) & 0xf;
+        }
+    }
+}

ARMeilleure/Decoders/OpCode32SimdMovn.cs

@@ -0,0 +1,12 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCode32SimdMovn : OpCode32Simd
+    {
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCode32SimdMovn(inst, address, opCode);
+
+        public OpCode32SimdMovn(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Size = (opCode >> 18) & 0x3;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT16Adr.cs

@@ -4,7 +4,6 @@ namespace ARMeilleure.Decoders
     {
         public int Rd { get; }
 
-        public bool Add => true;
         public int Immediate { get; }
 
         public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT16Adr(inst, address, opCode);

ARMeilleure/Decoders/OpCodeT16BImmCmp.cs

@@ -1,10 +1,10 @@
 namespace ARMeilleure.Decoders
 {
-    class OpCodeT16BImmCmp : OpCodeT16
+    class OpCodeT16BImmCmp : OpCodeT16, IOpCode32BImm
     {
         public int Rn { get; }
 
-        public int Immediate { get; }
+        public long Immediate { get; }
 
         public static new OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT16BImmCmp(inst, address, opCode);
 

ARMeilleure/Decoders/OpCodeT16IfThen.cs

@@ -1,5 +1,4 @@
 using System.Collections.Generic;
-using System.Reflection.Emit;
 
 namespace ARMeilleure.Decoders
 {

ARMeilleure/Decoders/OpCodeT32AluImm12.cs

@@ -0,0 +1,16 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32AluImm12 : OpCodeT32Alu, IOpCode32AluImm
+    {
+        public int Immediate { get; }
+
+        public bool IsRotated => false;
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32AluImm12(inst, address, opCode);
+
+        public OpCodeT32AluImm12(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Immediate = (opCode & 0xff) | ((opCode >> 4) & 0x700) | ((opCode >> 15) & 0x800);
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32AluReg.cs

@@ -0,0 +1,14 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32AluReg : OpCodeT32Alu, IOpCode32AluReg
+    {
+        public int Rm { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32AluReg(inst, address, opCode);
+
+        public OpCodeT32AluReg(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Rm = (opCode >> 0) & 0xf;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32BImm20.cs

@@ -1,6 +1,4 @@
-using ARMeilleure.Instructions;
-
-namespace ARMeilleure.Decoders
+namespace ARMeilleure.Decoders
 {
     class OpCodeT32BImm20 : OpCodeT32, IOpCode32BImm
     {

ARMeilleure/Decoders/OpCodeT32MemImm8D.cs

@@ -0,0 +1,31 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32MemImm8D : OpCodeT32, IOpCode32Mem
+    {
+        public int Rt { get; }
+        public int Rt2 { get; }
+        public int Rn { get; }
+        public bool WBack { get; }
+        public bool IsLoad { get; }
+        public bool Index { get; }
+        public bool Add { get; }
+        public int Immediate { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32MemImm8D(inst, address, opCode);
+
+        public OpCodeT32MemImm8D(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Rt2 = (opCode >> 8) & 0xf;
+            Rt = (opCode >> 12) & 0xf;
+            Rn = (opCode >> 16) & 0xf;
+
+            Index = ((opCode >> 24) & 1) != 0;
+            Add   = ((opCode >> 23) & 1) != 0;
+            WBack = ((opCode >> 21) & 1) != 0;
+
+            Immediate = (opCode & 0xff) << 2;
+
+            IsLoad = ((opCode >> 20) & 1) != 0;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32MemLdEx.cs

@@ -0,0 +1,24 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32MemLdEx : OpCodeT32, IOpCode32MemEx
+    {
+        public int Rd => 0;
+        public int Rt { get; }
+        public int Rn { get; }
+
+        public bool WBack => false;
+        public bool IsLoad => true;
+        public bool Index => false;
+        public bool Add => false;
+
+        public int Immediate => 0;
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32MemLdEx(inst, address, opCode);
+
+        public OpCodeT32MemLdEx(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Rt = (opCode >> 12) & 0xf;
+            Rn = (opCode >> 16) & 0xf;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32MemMult.cs

@@ -0,0 +1,52 @@
+using System.Numerics;
+
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32MemMult : OpCodeT32, IOpCode32MemMult
+    {
+        public int Rn { get; }
+
+        public int RegisterMask { get; }
+        public int Offset       { get; }
+        public int PostOffset   { get; }
+
+        public bool IsLoad { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32MemMult(inst, address, opCode);
+
+        public OpCodeT32MemMult(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Rn = (opCode >> 16) & 0xf;
+
+            bool isLoad = (opCode & (1 << 20)) != 0;
+            bool w      = (opCode & (1 << 21)) != 0;
+            bool u      = (opCode & (1 << 23)) != 0;
+            bool p      = (opCode & (1 << 24)) != 0;
+
+            RegisterMask = opCode & 0xffff;
+
+            int regsSize = BitOperations.PopCount((uint)RegisterMask) * 4;
+
+            if (!u)
+            {
+                Offset -= regsSize;
+            }
+
+            if (u == p)
+            {
+                Offset += 4;
+            }
+
+            if (w)
+            {
+                PostOffset = u ? regsSize : -regsSize;
+            }
+            else
+            {
+                PostOffset = 0;
+            }
+
+            IsLoad = isLoad;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32MemRsImm.cs

@@ -0,0 +1,30 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32MemRsImm : OpCodeT32, IOpCode32MemRsImm
+    {
+        public int Rt { get; }
+        public int Rn { get; }
+        public int Rm { get; }
+        public ShiftType ShiftType => ShiftType.Lsl;
+
+        public bool WBack => false;
+        public bool IsLoad { get; }
+        public bool Index => true;
+        public bool Add => true;
+
+        public int Immediate { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32MemRsImm(inst, address, opCode);
+
+        public OpCodeT32MemRsImm(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Rm = (opCode >> 0) & 0xf;
+            Rt = (opCode >> 12) & 0xf;
+            Rn = (opCode >> 16) & 0xf;
+
+            IsLoad = (opCode & (1 << 20)) != 0;
+
+            Immediate = (opCode >> 4) & 3;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32MemStEx.cs

@@ -0,0 +1,25 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32MemStEx : OpCodeT32, IOpCode32MemEx
+    {
+        public int Rd { get; }
+        public int Rt { get; }
+        public int Rn { get; }
+
+        public bool WBack => false;
+        public bool IsLoad => false;
+        public bool Index => false;
+        public bool Add => false;
+
+        public int Immediate => 0;
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32MemStEx(inst, address, opCode);
+
+        public OpCodeT32MemStEx(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Rd = (opCode >> 0) & 0xf;
+            Rt = (opCode >> 12) & 0xf;
+            Rn = (opCode >> 16) & 0xf;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32MovImm16.cs

@@ -0,0 +1,19 @@
+using ARMeilleure.Common;
+using System.Runtime.Intrinsics;
+
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32MovImm16 : OpCodeT32Alu, IOpCode32AluImm
+    {
+        public int Immediate { get; }
+
+        public bool IsRotated => false;
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32MovImm16(inst, address, opCode);
+
+        public OpCodeT32MovImm16(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Immediate = (opCode & 0xff) | ((opCode >> 4) & 0x700) | ((opCode >> 15) & 0x800) | ((opCode >> 4) & 0xf000);
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32ShiftReg.cs

@@ -0,0 +1,19 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32ShiftReg : OpCodeT32Alu, IOpCode32AluRsReg
+    {
+        public int Rm => Rn;
+        public int Rs { get; }
+
+        public ShiftType ShiftType { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32ShiftReg(inst, address, opCode);
+
+        public OpCodeT32ShiftReg(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Rs = (opCode >> 0) & 0xf;
+
+            ShiftType = (ShiftType)((opCode >> 21) & 3);
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32Tb.cs

@@ -0,0 +1,16 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32Tb : OpCodeT32, IOpCode32BReg
+    {
+        public int Rm { get; }
+        public int Rn { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32Tb(inst, address, opCode);
+
+        public OpCodeT32Tb(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Rm = (opCode >> 0) & 0xf;
+            Rn = (opCode >> 16) & 0xf;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeTable.cs

@@ -704,6 +704,7 @@ namespace ARMeilleure.Decoders
             SetA32("<<<<00110100xxxxxxxxxxxxxxxxxxxx", InstName.Movt,    InstEmit32.Movt,    OpCode32AluImm16.Create);
             SetA32("<<<<1110xxx1xxxxxxxx111xxxx1xxxx", InstName.Mrc,     InstEmit32.Mrc,     OpCode32System.Create);
             SetA32("<<<<11000101xxxxxxxx111xxxxxxxxx", InstName.Mrrc,    InstEmit32.Mrrc,    OpCode32System.Create);
+            SetA32("<<<<00010x001111xxxx000000000000", InstName.Mrs,     InstEmit32.Mrs,     OpCode32Mrs.Create);
             SetA32("<<<<00010x10xxxx111100000000xxxx", InstName.Msr,     InstEmit32.Msr,     OpCode32MsrReg.Create);
             SetA32("<<<<0000000xxxxx0000xxxx1001xxxx", InstName.Mul,     InstEmit32.Mul,     OpCode32AluMla.Create);
             SetA32("<<<<0011111x0000xxxxxxxxxxxxxxxx", InstName.Mvn,     InstEmit32.Mvn,     OpCode32AluImm.Create);
@@ -796,6 +797,10 @@ namespace ARMeilleure.Decoders
             SetA32("111100111x110000xxx0001100x0xxx0", InstName.Aese_V,      InstEmit32.Aese_V,      OpCode32Simd.Create);
             SetA32("111100111x110000xxx0001111x0xxx0", InstName.Aesimc_V,    InstEmit32.Aesimc_V,    OpCode32Simd.Create);
             SetA32("111100111x110000xxx0001110x0xxx0", InstName.Aesmc_V,     InstEmit32.Aesmc_V,     OpCode32Simd.Create);
+            SetA32("111100110x00xxx0xxx01100x1x0xxx0", InstName.Sha256h_V,   InstEmit32.Sha256h_V,   OpCode32SimdReg.Create);
+            SetA32("111100110x01xxx0xxx01100x1x0xxx0", InstName.Sha256h2_V,  InstEmit32.Sha256h2_V,  OpCode32SimdReg.Create);
+            SetA32("111100111x111010xxx0001111x0xxx0", InstName.Sha256su0_V, InstEmit32.Sha256su0_V, OpCode32Simd.Create);
+            SetA32("111100110x10xxx0xxx01100x1x0xxx0", InstName.Sha256su1_V, InstEmit32.Sha256su1_V, OpCode32SimdReg.Create);
             SetA32("1111001x0x<<xxxxxxxx0111xxx0xxxx", InstName.Vabd,        InstEmit32.Vabd_I,      OpCode32SimdReg.Create);
             SetA32("1111001x1x<<xxxxxxxx0111x0x0xxxx", InstName.Vabdl,       InstEmit32.Vabdl_I,     OpCode32SimdRegLong.Create);
             SetA32("<<<<11101x110000xxxx101x11x0xxxx", InstName.Vabs,        InstEmit32.Vabs_S,      OpCode32SimdS.Create);
@@ -804,8 +809,8 @@ namespace ARMeilleure.Decoders
             SetA32("111100100xxxxxxxxxxx1000xxx0xxxx", InstName.Vadd,        InstEmit32.Vadd_I,      OpCode32SimdReg.Create);
             SetA32("<<<<11100x11xxxxxxxx101xx0x0xxxx", InstName.Vadd,        InstEmit32.Vadd_S,      OpCode32SimdRegS.Create);
             SetA32("111100100x00xxxxxxxx1101xxx0xxxx", InstName.Vadd,        InstEmit32.Vadd_V,      OpCode32SimdReg.Create);
-            SetA32("1111001x1x<<xxxxxxxx0000x0x0xxxx", InstName.Vaddl,    InstEmit32.Vaddl_I,  OpCode32SimdRegLong.Create);
-            SetA32("1111001x1x<<xxxxxxxx0001x0x0xxxx", InstName.Vaddw,    InstEmit32.Vaddw_I,  OpCode32SimdRegWide.Create);
+            SetA32("1111001x1x<<xxxxxxx00000x0x0xxxx", InstName.Vaddl,       InstEmit32.Vaddl_I,     OpCode32SimdRegLong.Create);
+            SetA32("1111001x1x<<xxxxxxx00001x0x0xxxx", InstName.Vaddw,       InstEmit32.Vaddw_I,     OpCode32SimdRegWide.Create);
             SetA32("111100100x00xxxxxxxx0001xxx1xxxx", InstName.Vand,        InstEmit32.Vand_I,      OpCode32SimdBinary.Create);
             SetA32("111100100x01xxxxxxxx0001xxx1xxxx", InstName.Vbic,        InstEmit32.Vbic_I,      OpCode32SimdBinary.Create);
             SetA32("1111001x1x000xxxxxxx<<x10x11xxxx", InstName.Vbic,        InstEmit32.Vbic_II,     OpCode32SimdImm.Create);
@@ -874,6 +879,7 @@ namespace ARMeilleure.Decoders
             SetA32("111100100xxxxxxxxxxx1001xxx0xxxx", InstName.Vmla,        InstEmit32.Vmla_I,      OpCode32SimdReg.Create);
             SetA32("<<<<11100x00xxxxxxxx101xx0x0xxxx", InstName.Vmla,        InstEmit32.Vmla_S,      OpCode32SimdRegS.Create);
             SetA32("111100100x00xxxxxxxx1101xxx1xxxx", InstName.Vmla,        InstEmit32.Vmla_V,      OpCode32SimdReg.Create);
+            SetA32("1111001x1x<<xxxxxxx01000x0x0xxxx", InstName.Vmlal,       InstEmit32.Vmlal_I,     OpCode32SimdRegLong.Create);
             SetA32("1111001x1x<<xxxxxxxx010xx1x0xxxx", InstName.Vmls,        InstEmit32.Vmls_1,      OpCode32SimdRegElem.Create);
             SetA32("<<<<11100x00xxxxxxxx101xx1x0xxxx", InstName.Vmls,        InstEmit32.Vmls_S,      OpCode32SimdRegS.Create);
             SetA32("111100100x10xxxxxxxx1101xxx1xxxx", InstName.Vmls,        InstEmit32.Vmls_V,      OpCode32SimdReg.Create);
@@ -893,7 +899,7 @@ namespace ARMeilleure.Decoders
             SetA32("1111001x1x001000xxx0101000x1xxxx", InstName.Vmovl,       InstEmit32.Vmovl,       OpCode32SimdLong.Create);
             SetA32("1111001x1x010000xxx0101000x1xxxx", InstName.Vmovl,       InstEmit32.Vmovl,       OpCode32SimdLong.Create);
             SetA32("1111001x1x100000xxx0101000x1xxxx", InstName.Vmovl,       InstEmit32.Vmovl,       OpCode32SimdLong.Create);
-            SetA32("111100111x11xx10xxxx001000x0xxx0", InstName.Vmovn,    InstEmit32.Vmovn,    OpCode32SimdCmpZ.Create);
+            SetA32("111100111x11<<10xxxx001000x0xxx0", InstName.Vmovn,       InstEmit32.Vmovn,       OpCode32SimdMovn.Create);
             SetA32("<<<<11101111xxxxxxxx101000010000", InstName.Vmrs,        InstEmit32.Vmrs,        OpCode32SimdSpecial.Create);
             SetA32("<<<<11101110xxxxxxxx101000010000", InstName.Vmsr,        InstEmit32.Vmsr,        OpCode32SimdSpecial.Create);
             SetA32("1111001x1x<<xxxxxxxx100xx1x0xxxx", InstName.Vmul,        InstEmit32.Vmul_1,      OpCode32SimdRegElem.Create);
@@ -919,22 +925,33 @@ namespace ARMeilleure.Decoders
             SetA32("1111001x1x000xxxxxxx<<x10x01xxxx", InstName.Vorr,        InstEmit32.Vorr_II,     OpCode32SimdImm.Create);
             SetA32("111100100x<<xxxxxxxx1011x0x1xxxx", InstName.Vpadd,       InstEmit32.Vpadd_I,     OpCode32SimdReg.Create);
             SetA32("111100110x00xxxxxxxx1101x0x0xxxx", InstName.Vpadd,       InstEmit32.Vpadd_V,     OpCode32SimdReg.Create);
+            SetA32("111100111x11<<00xxxx0010xxx0xxxx", InstName.Vpaddl,      InstEmit32.Vpaddl,      OpCode32SimdCmpZ.Create);
             SetA32("1111001x0x<<xxxxxxxx1010x0x0xxxx", InstName.Vpmax,       InstEmit32.Vpmax_I,     OpCode32SimdReg.Create);
             SetA32("111100110x00xxxxxxxx1111x0x0xxxx", InstName.Vpmax,       InstEmit32.Vpmax_V,     OpCode32SimdReg.Create);
             SetA32("1111001x0x<<xxxxxxxx1010x0x1xxxx", InstName.Vpmin,       InstEmit32.Vpmin_I,     OpCode32SimdReg.Create);
             SetA32("111100110x10xxxxxxxx1111x0x0xxxx", InstName.Vpmin,       InstEmit32.Vpmin_V,     OpCode32SimdReg.Create);
+            SetA32("1111001x0xxxxxxxxxxx0000xxx1xxxx", InstName.Vqadd,       InstEmit32.Vqadd,       OpCode32SimdReg.Create);
+            SetA32("111100100x01xxxxxxxx1011xxx0xxxx", InstName.Vqdmulh,     InstEmit32.Vqdmulh,     OpCode32SimdReg.Create);
+            SetA32("111100100x10xxxxxxxx1011xxx0xxxx", InstName.Vqdmulh,     InstEmit32.Vqdmulh,     OpCode32SimdReg.Create);
+            SetA32("111100111x11<<10xxxx00101xx0xxx0", InstName.Vqmovn,      InstEmit32.Vqmovn,      OpCode32SimdMovn.Create);
+            SetA32("111100111x11<<10xxxx001001x0xxx0", InstName.Vqmovun,     InstEmit32.Vqmovun,     OpCode32SimdMovn.Create);
             SetA32("1111001x1x>>>xxxxxxx100101x1xxx0", InstName.Vqrshrn,     InstEmit32.Vqrshrn,     OpCode32SimdShImmNarrow.Create);
             SetA32("111100111x>>>xxxxxxx100001x1xxx0", InstName.Vqrshrun,    InstEmit32.Vqrshrun,    OpCode32SimdShImmNarrow.Create);
             SetA32("1111001x1x>>>xxxxxxx100100x1xxx0", InstName.Vqshrn,      InstEmit32.Vqshrn,      OpCode32SimdShImmNarrow.Create);
+            SetA32("111100111x>>>xxxxxxx100000x1xxx0", InstName.Vqshrun,     InstEmit32.Vqshrun,     OpCode32SimdShImmNarrow.Create);
+            SetA32("1111001x0xxxxxxxxxxx0010xxx1xxxx", InstName.Vqsub,       InstEmit32.Vqsub,       OpCode32SimdReg.Create);
             SetA32("111100111x111011xxxx010x0xx0xxxx", InstName.Vrecpe,      InstEmit32.Vrecpe,      OpCode32SimdSqrte.Create);
             SetA32("111100100x00xxxxxxxx1111xxx1xxxx", InstName.Vrecps,      InstEmit32.Vrecps,      OpCode32SimdReg.Create);
             SetA32("111100111x11xx00xxxx000<<xx0xxxx", InstName.Vrev,        InstEmit32.Vrev,        OpCode32SimdRev.Create);
+            SetA32("1111001x0x<<xxxxxxxx0001xxx0xxxx", InstName.Vrhadd,      InstEmit32.Vrhadd,      OpCode32SimdReg.Create);
             SetA32("111111101x1110xxxxxx101x01x0xxxx", InstName.Vrint,       InstEmit32.Vrint_RM,    OpCode32SimdS.Create);
             SetA32("<<<<11101x110110xxxx101x11x0xxxx", InstName.Vrint,       InstEmit32.Vrint_Z,     OpCode32SimdS.Create);
             SetA32("<<<<11101x110111xxxx101x01x0xxxx", InstName.Vrintx,      InstEmit32.Vrintx_S,    OpCode32SimdS.Create);
             SetA32("1111001x1x>>>xxxxxxx0010>xx1xxxx", InstName.Vrshr,       InstEmit32.Vrshr,       OpCode32SimdShImm.Create);
+            SetA32("111100101x>>>xxxxxxx100001x1xxx0", InstName.Vrshrn,      InstEmit32.Vrshrn,      OpCode32SimdShImmNarrow.Create);
             SetA32("111100111x111011xxxx010x1xx0xxxx", InstName.Vrsqrte,     InstEmit32.Vrsqrte,     OpCode32SimdSqrte.Create);
             SetA32("111100100x10xxxxxxxx1111xxx1xxxx", InstName.Vrsqrts,     InstEmit32.Vrsqrts,     OpCode32SimdReg.Create);
+            SetA32("1111001x1x>>>xxxxxxx0011>xx1xxxx", InstName.Vrsra,       InstEmit32.Vrsra,       OpCode32SimdShImm.Create);
             SetA32("111111100xxxxxxxxxxx101xx0x0xxxx", InstName.Vsel,        InstEmit32.Vsel,        OpCode32SimdSel.Create);
             SetA32("111100101x>>>xxxxxxx0101>xx1xxxx", InstName.Vshl,        InstEmit32.Vshl,        OpCode32SimdShImm.Create);
             SetA32("1111001x0xxxxxxxxxxx0100xxx0xxxx", InstName.Vshl,        InstEmit32.Vshl_I,      OpCode32SimdReg.Create);
@@ -965,7 +982,8 @@ namespace ARMeilleure.Decoders
             SetA32("111100110xxxxxxxxxxx1000xxx0xxxx", InstName.Vsub,        InstEmit32.Vsub_I,      OpCode32SimdReg.Create);
             SetA32("<<<<11100x11xxxxxxxx101xx1x0xxxx", InstName.Vsub,        InstEmit32.Vsub_S,      OpCode32SimdRegS.Create);
             SetA32("111100100x10xxxxxxxx1101xxx0xxxx", InstName.Vsub,        InstEmit32.Vsub_V,      OpCode32SimdReg.Create);
-            SetA32("1111001x1x<<xxxxxxxx0011x0x0xxxx", InstName.Vsubw,    InstEmit32.Vsubw_I,  OpCode32SimdRegWide.Create);
+            SetA32("1111001x1x<<xxxxxxx00010x0x0xxxx", InstName.Vsubl,       InstEmit32.Vsubl_I,     OpCode32SimdRegLong.Create);
+            SetA32("1111001x1x<<xxxxxxx00011x0x0xxxx", InstName.Vsubw,       InstEmit32.Vsubw_I,     OpCode32SimdRegWide.Create);
             SetA32("111100111x11xxxxxxxx10xxxxx0xxxx", InstName.Vtbl,        InstEmit32.Vtbl,        OpCode32SimdTbl.Create);
             SetA32("111100111x11<<10xxxx00001xx0xxxx", InstName.Vtrn,        InstEmit32.Vtrn,        OpCode32SimdCmpZ.Create);
             SetA32("111100100x<<xxxxxxxx1000xxx1xxxx", InstName.Vtst,        InstEmit32.Vtst,        OpCode32SimdReg.Create);
@@ -1003,7 +1021,7 @@ namespace ARMeilleure.Decoders
             SetT16("01000101xxxxxxxx", InstName.Cmp,    InstEmit32.Cmp,     OpCodeT16AluRegHigh.Create);
             SetT16("01000110xxxxxxxx", InstName.Mov,    InstEmit32.Mov,     OpCodeT16AluRegHigh.Create);
             SetT16("010001110xxxx000", InstName.Bx,     InstEmit32.Bx,      OpCodeT16BReg.Create);
-            SetT16("010001111xxxx000", InstName.Blx,    InstEmit32.Blx,     OpCodeT16BReg.Create);
+            SetT16("010001111xxxx000", InstName.Blx,    InstEmit32.Blxr,    OpCodeT16BReg.Create);
             SetT16("01001xxxxxxxxxxx", InstName.Ldr,    InstEmit32.Ldr,     OpCodeT16MemLit.Create);
             SetT16("0101000xxxxxxxxx", InstName.Str,    InstEmit32.Str,     OpCodeT16MemReg.Create);
             SetT16("0101001xxxxxxxxx", InstName.Strh,   InstEmit32.Strh,    OpCodeT16MemReg.Create);
@@ -1051,6 +1069,7 @@ namespace ARMeilleure.Decoders
             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("11110x100000xxxx0xxxxxxxxxxxxxxx", InstName.Add,      InstEmit32.Add,      OpCodeT32AluImm12.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);
@@ -1059,20 +1078,28 @@ namespace ARMeilleure.Decoders
             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("111110101011xxxx1111xxxx1000xxxx", InstName.Clz,      InstEmit32.Clz,      OpCodeT32AluReg.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("111010001101xxxxxxxx111110101111", InstName.Lda,      InstEmit32.Lda,      OpCodeT32MemLdEx.Create);
+            SetT32("111010001101xxxxxxxx111111101111", InstName.Ldaex,    InstEmit32.Ldaex,    OpCodeT32MemLdEx.Create);
+            SetT32("1110100010x1xxxxxxxxxxxxxxxxxxxx", InstName.Ldm,      InstEmit32.Ldm,      OpCodeT32MemMult.Create);
+            SetT32("1110100100x1xxxxxxxxxxxxxxxxxxxx", InstName.Ldm,      InstEmit32.Ldm,      OpCodeT32MemMult.Create);
             SetT32("111110000101xxxx<<<<10x1xxxxxxxx", InstName.Ldr,      InstEmit32.Ldr,      OpCodeT32MemImm8.Create);
             SetT32("111110000101xxxx<<<<1100xxxxxxxx", InstName.Ldr,      InstEmit32.Ldr,      OpCodeT32MemImm8.Create);
             SetT32("111110000101xxxx<<<<11x1xxxxxxxx", InstName.Ldr,      InstEmit32.Ldr,      OpCodeT32MemImm8.Create);
             SetT32("111110001101xxxxxxxxxxxxxxxxxxxx", InstName.Ldr,      InstEmit32.Ldr,      OpCodeT32MemImm12.Create);
+            SetT32("111110000101<<<<xxxx000000xxxxxx", InstName.Ldr,      InstEmit32.Ldr,      OpCodeT32MemRsImm.Create);
             SetT32("111110000001xxxx<<<<10x1xxxxxxxx", InstName.Ldrb,     InstEmit32.Ldrb,     OpCodeT32MemImm8.Create);
             SetT32("111110000001xxxx<<<<1100xxxxxxxx", InstName.Ldrb,     InstEmit32.Ldrb,     OpCodeT32MemImm8.Create);
             SetT32("111110000001xxxx<<<<11x1xxxxxxxx", InstName.Ldrb,     InstEmit32.Ldrb,     OpCodeT32MemImm8.Create);
             SetT32("111110001001xxxxxxxxxxxxxxxxxxxx", InstName.Ldrb,     InstEmit32.Ldrb,     OpCodeT32MemImm12.Create);
+            SetT32("11101000x111<<<<xxxxxxxxxxxxxxxx", InstName.Ldrd,     InstEmit32.Ldrd,     OpCodeT32MemImm8D.Create);
+            SetT32("11101001x1x1<<<<xxxxxxxxxxxxxxxx", InstName.Ldrd,     InstEmit32.Ldrd,     OpCodeT32MemImm8D.Create);
             SetT32("111110000011xxxx<<<<10x1xxxxxxxx", InstName.Ldrh,     InstEmit32.Ldrh,     OpCodeT32MemImm8.Create);
             SetT32("111110000011xxxx<<<<1100xxxxxxxx", InstName.Ldrh,     InstEmit32.Ldrh,     OpCodeT32MemImm8.Create);
             SetT32("111110000011xxxx<<<<11x1xxxxxxxx", InstName.Ldrh,     InstEmit32.Ldrh,     OpCodeT32MemImm8.Create);
@@ -1086,9 +1113,12 @@ namespace ARMeilleure.Decoders
             SetT32("111110010011xxxx<<<<11x1xxxxxxxx", InstName.Ldrsh,    InstEmit32.Ldrsh,    OpCodeT32MemImm8.Create);
             SetT32("111110011011xxxxxxxxxxxxxxxxxxxx", InstName.Ldrsh,    InstEmit32.Ldrsh,    OpCodeT32MemImm12.Create);
             SetT32("11101010010x11110xxxxxxxxxxxxxxx", InstName.Mov,      InstEmit32.Mov,      OpCodeT32AluRsImm.Create);
+            SetT32("111110100xxxxxxx1111xxxx0000xxxx", InstName.Mov,      InstEmit32.Mov,      OpCodeT32ShiftReg.Create);
             SetT32("11110x00010x11110xxxxxxxxxxxxxxx", InstName.Mov,      InstEmit32.Mov,      OpCodeT32AluImm.Create);
+            SetT32("11110x100100xxxx0xxxxxxxxxxxxxxx", InstName.Mov,      InstEmit32.Mov,      OpCodeT32MovImm16.Create);
             SetT32("11101010011x11110xxxxxxxxxxxxxxx", InstName.Mvn,      InstEmit32.Mvn,      OpCodeT32AluRsImm.Create);
             SetT32("11110x00011x11110xxxxxxxxxxxxxxx", InstName.Mvn,      InstEmit32.Mvn,      OpCodeT32AluImm.Create);
+            SetT32("11110011101011111000000000000000", InstName.Nop,      InstEmit32.Nop,      OpCodeT32.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);
@@ -1097,14 +1127,22 @@ namespace ARMeilleure.Decoders
             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("111010001100xxxxxxxx11111110xxxx", InstName.Stlex,    InstEmit32.Stlex,    OpCodeT32MemStEx.Create);
+            SetT32("1110100010x0xxxx0xxxxxxxxxxxxxxx", InstName.Stm,      InstEmit32.Stm,      OpCodeT32MemMult.Create);
+            SetT32("1110100100x0xxxx0xxxxxxxxxxxxxxx", InstName.Stm,      InstEmit32.Stm,      OpCodeT32MemMult.Create);
             SetT32("111110000100xxxxxxxx1<<>xxxxxxxx", InstName.Str,      InstEmit32.Str,      OpCodeT32MemImm8.Create);
             SetT32("111110001100xxxxxxxxxxxxxxxxxxxx", InstName.Str,      InstEmit32.Str,      OpCodeT32MemImm12.Create);
+            SetT32("111110000100<<<<xxxx000000xxxxxx", InstName.Str,      InstEmit32.Str,      OpCodeT32MemRsImm.Create);
             SetT32("111110000000xxxxxxxx1<<>xxxxxxxx", InstName.Strb,     InstEmit32.Strb,     OpCodeT32MemImm8.Create);
             SetT32("111110001000xxxxxxxxxxxxxxxxxxxx", InstName.Strb,     InstEmit32.Strb,     OpCodeT32MemImm12.Create);
+            SetT32("11101000x110<<<<xxxxxxxxxxxxxxxx", InstName.Strd,     InstEmit32.Strd,     OpCodeT32MemImm8D.Create);
+            SetT32("11101001x1x0<<<<xxxxxxxxxxxxxxxx", InstName.Strd,     InstEmit32.Strd,     OpCodeT32MemImm8D.Create);
             SetT32("111110000010xxxxxxxx1<<>xxxxxxxx", InstName.Strh,     InstEmit32.Strh,     OpCodeT32MemImm8.Create);
             SetT32("111110001010xxxxxxxxxxxxxxxxxxxx", InstName.Strh,     InstEmit32.Strh,     OpCodeT32MemImm12.Create);
             SetT32("11101011101<xxxx0xxx<<<<xxxxxxxx", InstName.Sub,      InstEmit32.Sub,      OpCodeT32AluRsImm.Create);
             SetT32("11110x01101<xxxx0xxx<<<<xxxxxxxx", InstName.Sub,      InstEmit32.Sub,      OpCodeT32AluImm.Create);
+            SetT32("111010001101xxxx111100000000xxxx", InstName.Tbb,      InstEmit32.Tbb,      OpCodeT32Tb.Create);
+            SetT32("111010001101xxxx111100000001xxxx", InstName.Tbh,      InstEmit32.Tbh,      OpCodeT32Tb.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);

ARMeilleure/Diagnostics/EventSources/AddressTableEventSource.cs

@@ -1,51 +0,0 @@
-using System.Diagnostics.Tracing;
-
-namespace ARMeilleure.Diagnostics.EventSources
-{
-    [EventSource(Name = "ARMeilleure")]
-    class AddressTableEventSource : EventSource
-    {
-        public static readonly AddressTableEventSource Log = new();
-
-        private ulong _size;
-        private ulong _leafSize;
-        private PollingCounter _sizeCounter;
-        private PollingCounter _leafSizeCounter;
-
-        public AddressTableEventSource()
-        {
-            _sizeCounter = new PollingCounter("addr-tab-alloc", this, () => _size / 1024d / 1024d)
-            {
-                DisplayName = "AddressTable Total Bytes Allocated",
-                DisplayUnits = "MB"
-            };
-
-            _leafSizeCounter = new PollingCounter("addr-tab-leaf-alloc", this, () => _leafSize / 1024d / 1024d)
-            {
-                DisplayName = "AddressTable Total Leaf Bytes Allocated",
-                DisplayUnits = "MB"
-            };
-        }
-
-        public void Allocated(int bytes, bool leaf)
-        {
-            _size += (uint)bytes;
-
-            if (leaf)
-            {
-                _leafSize += (uint)bytes;
-            }
-        }
-
-        protected override void Dispose(bool disposing)
-        {
-            _leafSizeCounter.Dispose();
-            _leafSizeCounter = null;
-
-            _sizeCounter.Dispose();
-            _sizeCounter = null;
-
-            base.Dispose(disposing);
-        }
-    }
-}

ARMeilleure/Diagnostics/TranslatorEventSource.cs

@@ -0,0 +1,67 @@
+using System.Diagnostics.Tracing;
+using System.Threading;
+
+namespace ARMeilleure.Diagnostics
+{
+    [EventSource(Name = "ARMeilleure")]
+    class TranslatorEventSource : EventSource
+    {
+        public static readonly TranslatorEventSource Log = new();
+
+        private int _rejitQueue;
+        private ulong _funcTabSize;
+        private ulong _funcTabLeafSize;
+        private PollingCounter _rejitQueueCounter;
+        private PollingCounter _funcTabSizeCounter;
+        private PollingCounter _funcTabLeafSizeCounter;
+
+        public TranslatorEventSource()
+        {
+            _rejitQueueCounter = new PollingCounter("rejit-queue-length", this, () => _rejitQueue)
+            {
+                DisplayName = "Rejit Queue Length"
+            };
+
+            _funcTabSizeCounter = new PollingCounter("addr-tab-alloc", this, () => _funcTabSize / 1024d / 1024d)
+            {
+                DisplayName = "AddressTable Total Bytes Allocated",
+                DisplayUnits = "MB"
+            };
+
+            _funcTabLeafSizeCounter = new PollingCounter("addr-tab-leaf-alloc", this, () => _funcTabLeafSize / 1024d / 1024d)
+            {
+                DisplayName = "AddressTable Total Leaf Bytes Allocated",
+                DisplayUnits = "MB"
+            };
+        }
+
+        public void RejitQueueAdd(int count)
+        {
+            Interlocked.Add(ref _rejitQueue, count);
+        }
+
+        public void AddressTableAllocated(int bytes, bool leaf)
+        {
+            _funcTabSize += (uint)bytes;
+
+            if (leaf)
+            {
+                _funcTabLeafSize += (uint)bytes;
+            }
+        }
+
+        protected override void Dispose(bool disposing)
+        {
+            _rejitQueueCounter.Dispose();
+            _rejitQueueCounter = null;
+
+            _funcTabLeafSizeCounter.Dispose();
+            _funcTabLeafSizeCounter = null;
+
+            _funcTabSizeCounter.Dispose();
+            _funcTabSizeCounter = null;
+
+            base.Dispose(disposing);
+        }
+    }
+}

ARMeilleure/Instructions/InstEmitException32.cs

@@ -1,7 +1,5 @@
 using ARMeilleure.Decoders;
 using ARMeilleure.Translation;
-
-using static ARMeilleure.Instructions.InstEmitFlowHelper;
 using static ARMeilleure.IntermediateRepresentation.Operand.Factory;
 
 namespace ARMeilleure.Instructions

ARMeilleure/Instructions/InstEmitFlow32.cs

@@ -88,7 +88,7 @@ namespace ARMeilleure.Instructions
         {
             OpCodeT16BImmCmp op = (OpCodeT16BImmCmp)context.CurrOp;
 
-            Operand value = GetIntOrZR(context, op.Rn);
+            Operand value = GetIntA32(context, op.Rn);
             Operand lblTarget = context.GetLabel((ulong)op.Immediate);
 
             if (onNotZero)
@@ -107,5 +107,30 @@ namespace ARMeilleure.Instructions
 
             context.SetIfThenBlockState(op.IfThenBlockConds);
         }
+
+        public static void Tbb(ArmEmitterContext context) => EmitTb(context, halfword: false);
+        public static void Tbh(ArmEmitterContext context)  => EmitTb(context, halfword: true);
+
+        private static void EmitTb(ArmEmitterContext context, bool halfword)
+        {
+            OpCodeT32Tb op = (OpCodeT32Tb)context.CurrOp;
+
+            Operand halfwords;
+
+            if (halfword)
+            {
+                Operand address = context.Add(GetIntA32(context, op.Rn), context.ShiftLeft(GetIntA32(context, op.Rm), Const(1)));
+                halfwords = InstEmitMemoryHelper.EmitReadInt(context, address, 1);
+            }
+            else
+            {
+                Operand address = context.Add(GetIntA32(context, op.Rn), GetIntA32(context, op.Rm));
+                halfwords = InstEmitMemoryHelper.EmitReadIntAligned(context, address, 0);
+            }
+
+            Operand targetAddress = context.Add(Const((int)op.GetPc()), context.ShiftLeft(halfwords, Const(1)));
+
+            EmitVirtualJump(context, targetAddress, isReturn: false);
+        }
     }
 }

ARMeilleure/Instructions/InstEmitMemory32.cs

@@ -205,13 +205,13 @@ namespace ARMeilleure.Instructions
                     context.BranchIfTrue(lblBigEndian, GetFlag(PState.EFlag));
 
                     Load(op.Rt, 0, WordSizeLog2);
-                    Load(op.Rt | 1, 4, WordSizeLog2);
+                    Load(op.Rt2, 4, WordSizeLog2);
 
                     context.Branch(lblEnd);
 
                     context.MarkLabel(lblBigEndian);
 
-                    Load(op.Rt | 1, 0, WordSizeLog2);
+                    Load(op.Rt2, 0, WordSizeLog2);
                     Load(op.Rt, 4, WordSizeLog2);
 
                     context.MarkLabel(lblEnd);
@@ -238,13 +238,13 @@ namespace ARMeilleure.Instructions
                     context.BranchIfTrue(lblBigEndian, GetFlag(PState.EFlag));
 
                     Store(op.Rt, 0, WordSizeLog2);
-                    Store(op.Rt | 1, 4, WordSizeLog2);
+                    Store(op.Rt2, 4, WordSizeLog2);
 
                     context.Branch(lblEnd);
 
                     context.MarkLabel(lblBigEndian);
 
-                    Store(op.Rt | 1, 0, WordSizeLog2);
+                    Store(op.Rt2, 0, WordSizeLog2);
                     Store(op.Rt, 4, WordSizeLog2);
 
                     context.MarkLabel(lblEnd);

ARMeilleure/Instructions/InstEmitMemoryHelper.cs

@@ -123,6 +123,41 @@ namespace ARMeilleure.Instructions
                    context.CurrOp is OpCodeSimdMemSs);
         }
 
+        public static Operand EmitReadInt(ArmEmitterContext context, Operand address, int size)
+        {
+            Operand temp = context.AllocateLocal(size == 3 ? OperandType.I64 : OperandType.I32);
+
+            Operand lblSlowPath = Label();
+            Operand lblEnd      = Label();
+
+            Operand physAddr = EmitPtPointerLoad(context, address, lblSlowPath, write: false, size);
+
+            Operand value = default;
+
+            switch (size)
+            {
+                case 0: value = context.Load8 (physAddr);                  break;
+                case 1: value = context.Load16(physAddr);                  break;
+                case 2: value = context.Load  (OperandType.I32, physAddr); break;
+                case 3: value = context.Load  (OperandType.I64, physAddr); break;
+            }
+
+            context.Copy(temp, value);
+
+            if (!context.Memory.Type.IsHostMapped())
+            {
+                context.Branch(lblEnd);
+
+                context.MarkLabel(lblSlowPath, BasicBlockFrequency.Cold);
+
+                context.Copy(temp, EmitReadIntFallback(context, address, size));
+
+                context.MarkLabel(lblEnd);
+            }
+
+            return temp;
+        }
+
         private static void EmitReadInt(ArmEmitterContext context, Operand address, int rt, int size)
         {
             Operand lblSlowPath = Label();
@@ -419,6 +454,11 @@ namespace ARMeilleure.Instructions
         }
 
         private static void EmitReadIntFallback(ArmEmitterContext context, Operand address, int rt, int size)
+        {
+            SetInt(context, rt, EmitReadIntFallback(context, address, size));
+        }
+
+        private static Operand EmitReadIntFallback(ArmEmitterContext context, Operand address, int size)
         {
             MethodInfo info = null;
 
@@ -430,7 +470,7 @@ namespace ARMeilleure.Instructions
                 case 3: info = typeof(NativeInterface).GetMethod(nameof(NativeInterface.ReadUInt64)); break;
             }
 
-            SetInt(context, rt, context.Call(info, address));
+            return context.Call(info, address);
         }
 
         private static void EmitReadVectorFallback(
@@ -547,7 +587,7 @@ namespace ARMeilleure.Instructions
         {
             switch (context.CurrOp)
             {
-                case OpCode32MemRsImm op: return GetMShiftedByImmediate(context, op, setCarry);
+                case IOpCode32MemRsImm op: return GetMShiftedByImmediate(context, op, setCarry);
 
                 case IOpCode32MemReg op: return GetIntA32(context, op.Rm);
 
@@ -564,7 +604,7 @@ namespace ARMeilleure.Instructions
             return new InvalidOperationException($"Invalid OpCode type \"{opCode?.GetType().Name ?? "null"}\".");
         }
 
-        public static Operand GetMShiftedByImmediate(ArmEmitterContext context, OpCode32MemRsImm op, bool setCarry)
+        public static Operand GetMShiftedByImmediate(ArmEmitterContext context, IOpCode32MemRsImm op, bool setCarry)
         {
             Operand m = GetIntA32(context, op.Rm);
 

ARMeilleure/Instructions/InstEmitSimdArithmetic32.cs

@@ -2,8 +2,6 @@
 using ARMeilleure.IntermediateRepresentation;
 using ARMeilleure.Translation;
 using System;
-using System.Diagnostics;
-
 using static ARMeilleure.Instructions.InstEmitFlowHelper;
 using static ARMeilleure.Instructions.InstEmitHelper;
 using static ARMeilleure.Instructions.InstEmitSimdHelper;
@@ -779,6 +777,13 @@ namespace ARMeilleure.Instructions
             }
         }
 
+        public static void Vmlal_I(ArmEmitterContext context)
+        {
+            OpCode32SimdReg op = (OpCode32SimdReg)context.CurrOp;
+
+            EmitVectorTernaryLongOpI32(context, (d, n, m) => context.Add(d, context.Multiply(n, m)), !op.U);
+        }
+
         public static void Vmls_S(ArmEmitterContext context)
         {
             if (Optimizations.FastFP && Optimizations.UseSse2)
@@ -994,6 +999,13 @@ namespace ARMeilleure.Instructions
             }
         }
 
+        public static void Vpaddl(ArmEmitterContext context)
+        {
+            OpCode32Simd op = (OpCode32Simd)context.CurrOp;
+
+            EmitVectorPairwiseLongOpI32(context, (op1, op2) => context.Add(op1, op2), (op.Opc & 1) == 0);
+        }
+
         public static void Vpmax_V(ArmEmitterContext context)
         {
             if (Optimizations.FastFP && Optimizations.UseSse2)
@@ -1054,6 +1066,62 @@ namespace ARMeilleure.Instructions
             }
         }
 
+        public static void Vqadd(ArmEmitterContext context)
+        {
+            OpCode32SimdReg op = (OpCode32SimdReg)context.CurrOp;
+
+            EmitSaturatingAddSubBinaryOp(context, add: true, !op.U);
+        }
+
+        public static void Vqdmulh(ArmEmitterContext context)
+        {
+            OpCode32SimdReg op = (OpCode32SimdReg)context.CurrOp;
+            int eSize = 8 << op.Size;
+
+            EmitVectorBinaryOpI32(context, (op1, op2) =>
+            {
+                if (op.Size == 2)
+                {
+                    op1 = context.SignExtend32(OperandType.I64, op1);
+                    op2 = context.SignExtend32(OperandType.I64, op2);
+                }
+
+                Operand res = context.Multiply(op1, op2);
+                res = context.ShiftRightSI(res, Const(eSize - 1));
+                res = EmitSatQ(context, res, eSize, signedSrc: true, signedDst: true);
+
+                if (op.Size == 2)
+                {
+                    res = context.ConvertI64ToI32(res);
+                }
+
+                return res;
+            }, signed: true);
+        }
+
+        public static void Vqmovn(ArmEmitterContext context)
+        {
+            OpCode32SimdMovn op = (OpCode32SimdMovn)context.CurrOp;
+
+            bool signed = !op.Q;
+
+            EmitVectorUnaryNarrowOp32(context, (op1) => EmitSatQ(context, op1, 8 << op.Size, signed, signed), signed);
+        }
+
+        public static void Vqmovun(ArmEmitterContext context)
+        {
+            OpCode32SimdMovn op = (OpCode32SimdMovn)context.CurrOp;
+
+            EmitVectorUnaryNarrowOp32(context, (op1) => EmitSatQ(context, op1, 8 << op.Size, signedSrc: true, signedDst: false), signed: true);
+        }
+
+        public static void Vqsub(ArmEmitterContext context)
+        {
+            OpCode32SimdReg op = (OpCode32SimdReg)context.CurrOp;
+
+            EmitSaturatingAddSubBinaryOp(context, add: false, !op.U);
+        }
+
         public static void Vrev(ArmEmitterContext context)
         {
             OpCode32SimdRev op = (OpCode32SimdRev)context.CurrOp;
@@ -1204,6 +1272,30 @@ namespace ARMeilleure.Instructions
             }
         }
 
+        public static void Vrhadd(ArmEmitterContext context)
+        {
+            OpCode32SimdReg op = (OpCode32SimdReg)context.CurrOp;
+
+            EmitVectorBinaryOpI32(context, (op1, op2) =>
+            {
+                if (op.Size == 2)
+                {
+                    op1 = context.ZeroExtend32(OperandType.I64, op1);
+                    op2 = context.ZeroExtend32(OperandType.I64, op2);
+                }
+
+                Operand res = context.Add(context.Add(op1, op2), Const(op1.Type, 1L));
+                res = context.ShiftRightUI(res, Const(1));
+
+                if (op.Size == 2)
+                {
+                    res = context.ConvertI64ToI32(res);
+                }
+
+                return res;
+            }, !op.U);
+        }
+
         public static void Vrsqrte(ArmEmitterContext context)
         {
             OpCode32SimdSqrte op = (OpCode32SimdSqrte)context.CurrOp;
@@ -1351,6 +1443,13 @@ namespace ARMeilleure.Instructions
             }
         }
 
+        public static void Vsubl_I(ArmEmitterContext context)
+        {
+            OpCode32SimdRegLong op = (OpCode32SimdRegLong)context.CurrOp;
+
+            EmitVectorBinaryLongOpI32(context, (op1, op2) => context.Subtract(op1, op2), !op.U);
+        }
+
         public static void Vsubw_I(ArmEmitterContext context)
         {
             OpCode32SimdRegWide op = (OpCode32SimdRegWide)context.CurrOp;
@@ -1358,6 +1457,46 @@ namespace ARMeilleure.Instructions
             EmitVectorBinaryWideOpI32(context, (op1, op2) => context.Subtract(op1, op2), !op.U);
         }
 
+        private static void EmitSaturatingAddSubBinaryOp(ArmEmitterContext context, bool add, bool signed)
+        {
+            OpCode32Simd op = (OpCode32Simd)context.CurrOp;
+
+            EmitVectorBinaryOpI32(context, (ne, me) =>
+            {
+                if (op.Size <= 2)
+                {
+                    if (op.Size == 2)
+                    {
+                        ne = signed ? context.SignExtend32(OperandType.I64, ne) : context.ZeroExtend32(OperandType.I64, ne);
+                        me = signed ? context.SignExtend32(OperandType.I64, me) : context.ZeroExtend32(OperandType.I64, me);
+                    }
+
+                    Operand res = add ? context.Add(ne, me) : context.Subtract(ne, me);
+
+                    res = EmitSatQ(context, res, 8 << op.Size, signedSrc: true, signed);
+
+                    if (op.Size == 2)
+                    {
+                        res = context.ConvertI64ToI32(res);
+                    }
+
+                    return res;
+                }
+                else if (add) /* if (op.Size == 3) */
+                {
+                    return signed
+                        ? EmitBinarySignedSatQAdd(context, ne, me)
+                        : EmitBinaryUnsignedSatQAdd(context, ne, me);
+                }
+                else /* if (sub) */
+                {
+                    return signed
+                        ? EmitBinarySignedSatQSub(context, ne, me)
+                        : EmitBinaryUnsignedSatQSub(context, ne, me);
+                }
+            }, signed);
+        }
+
         private static void EmitSse41MaxMinNumOpF32(ArmEmitterContext context, bool isMaxNum, bool scalar)
         {
             IOpCode32Simd op = (IOpCode32Simd)context.CurrOp;

ARMeilleure/Instructions/InstEmitSimdHash.cs

@@ -100,7 +100,7 @@ namespace ARMeilleure.Instructions
             Operand n = GetVec(op.Rn);
             Operand m = GetVec(op.Rm);
 
-            Operand res = context.Call(typeof(SoftFallback).GetMethod(nameof(SoftFallback.HashLower)), d, n, m);
+            Operand res = InstEmitSimdHashHelper.EmitSha256h(context, d, n, m, part2: false);
 
             context.Copy(GetVec(op.Rd), res);
         }
@@ -113,7 +113,7 @@ namespace ARMeilleure.Instructions
             Operand n = GetVec(op.Rn);
             Operand m = GetVec(op.Rm);
 
-            Operand res = context.Call(typeof(SoftFallback).GetMethod(nameof(SoftFallback.HashUpper)), d, n, m);
+            Operand res = InstEmitSimdHashHelper.EmitSha256h(context, n, d, m, part2: true);
 
             context.Copy(GetVec(op.Rd), res);
         }
@@ -125,7 +125,7 @@ namespace ARMeilleure.Instructions
             Operand d = GetVec(op.Rd);
             Operand n = GetVec(op.Rn);
 
-            Operand res = context.Call(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Sha256SchedulePart1)), d, n);
+            Operand res = InstEmitSimdHashHelper.EmitSha256su0(context, d, n);
 
             context.Copy(GetVec(op.Rd), res);
         }
@@ -138,7 +138,7 @@ namespace ARMeilleure.Instructions
             Operand n = GetVec(op.Rn);
             Operand m = GetVec(op.Rm);
 
-            Operand res = context.Call(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Sha256SchedulePart2)), d, n, m);
+            Operand res = InstEmitSimdHashHelper.EmitSha256su1(context, d, n, m);
 
             context.Copy(GetVec(op.Rd), res);
         }

ARMeilleure/Instructions/InstEmitSimdHash32.cs

@@ -0,0 +1,64 @@
+using ARMeilleure.Decoders;
+using ARMeilleure.IntermediateRepresentation;
+using ARMeilleure.Translation;
+
+using static ARMeilleure.Instructions.InstEmitHelper;
+
+namespace ARMeilleure.Instructions
+{
+    static partial class InstEmit32
+    {
+#region "Sha256"
+        public static void Sha256h_V(ArmEmitterContext context)
+        {
+            OpCode32SimdReg op = (OpCode32SimdReg)context.CurrOp;
+
+            Operand d = GetVecA32(op.Qd);
+            Operand n = GetVecA32(op.Qn);
+            Operand m = GetVecA32(op.Qm);
+
+            Operand res = InstEmitSimdHashHelper.EmitSha256h(context, d, n, m, part2: false);
+
+            context.Copy(GetVecA32(op.Qd), res);
+        }
+
+        public static void Sha256h2_V(ArmEmitterContext context)
+        {
+            OpCode32SimdReg op = (OpCode32SimdReg)context.CurrOp;
+
+            Operand d = GetVecA32(op.Qd);
+            Operand n = GetVecA32(op.Qn);
+            Operand m = GetVecA32(op.Qm);
+
+            Operand res = InstEmitSimdHashHelper.EmitSha256h(context, n, d, m, part2: true);
+
+            context.Copy(GetVecA32(op.Qd), res);
+        }
+
+        public static void Sha256su0_V(ArmEmitterContext context)
+        {
+            OpCode32Simd op = (OpCode32Simd)context.CurrOp;
+
+            Operand d = GetVecA32(op.Qd);
+            Operand m = GetVecA32(op.Qm);
+
+            Operand res = InstEmitSimdHashHelper.EmitSha256su0(context, d, m);
+
+            context.Copy(GetVecA32(op.Qd), res);
+        }
+
+        public static void Sha256su1_V(ArmEmitterContext context)
+        {
+            OpCode32SimdReg op = (OpCode32SimdReg)context.CurrOp;
+
+            Operand d = GetVecA32(op.Qd);
+            Operand n = GetVecA32(op.Qn);
+            Operand m = GetVecA32(op.Qm);
+
+            Operand res = InstEmitSimdHashHelper.EmitSha256su1(context, d, n, m);
+
+            context.Copy(GetVecA32(op.Qd), res);
+        }
+#endregion
+    }
+}

ARMeilleure/Instructions/InstEmitSimdHashHelper.cs

@@ -0,0 +1,56 @@
+using ARMeilleure.IntermediateRepresentation;
+using ARMeilleure.Translation;
+using System;
+
+using static ARMeilleure.IntermediateRepresentation.Operand.Factory;
+
+namespace ARMeilleure.Instructions
+{
+    static class InstEmitSimdHashHelper
+    {
+        public static Operand EmitSha256h(ArmEmitterContext context, Operand x, Operand y, Operand w, bool part2)
+        {
+            if (Optimizations.UseSha)
+            {
+                Operand src1 = context.AddIntrinsic(Intrinsic.X86Shufps, y, x, Const(0xbb));
+                Operand src2 = context.AddIntrinsic(Intrinsic.X86Shufps, y, x, Const(0x11));
+                Operand w2 = context.AddIntrinsic(Intrinsic.X86Punpckhqdq, w, w);
+
+                Operand round2 = context.AddIntrinsic(Intrinsic.X86Sha256Rnds2, src1, src2, w);
+                Operand round4 = context.AddIntrinsic(Intrinsic.X86Sha256Rnds2, src2, round2, w2);
+                
+                Operand res = context.AddIntrinsic(Intrinsic.X86Shufps, round4, round2, Const(part2 ? 0x11 : 0xbb));
+                
+                return res;
+            }
+
+            String method = part2 ? nameof(SoftFallback.HashUpper) : nameof(SoftFallback.HashLower);
+            return context.Call(typeof(SoftFallback).GetMethod(method), x, y, w);
+        }
+
+        public static Operand EmitSha256su0(ArmEmitterContext context, Operand x, Operand y)
+        {
+            if (Optimizations.UseSha)
+            {
+                return context.AddIntrinsic(Intrinsic.X86Sha256Msg1, x, y);
+            }
+
+            return context.Call(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Sha256SchedulePart1)), x, y);
+        }
+
+        public static Operand EmitSha256su1(ArmEmitterContext context, Operand x, Operand y, Operand z)
+        {
+            if (Optimizations.UseSha && Optimizations.UseSsse3)
+            {
+                Operand extr = context.AddIntrinsic(Intrinsic.X86Palignr, z, y, Const(4));
+                Operand tmp = context.AddIntrinsic(Intrinsic.X86Paddd, extr, x);
+
+                Operand res = context.AddIntrinsic(Intrinsic.X86Sha256Msg2, tmp, z);
+
+                return res;
+            }
+
+            return context.Call(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Sha256SchedulePart2)), x, y, z);
+        }
+    }
+}

ARMeilleure/Instructions/InstEmitSimdHelper.cs

@@ -1352,7 +1352,7 @@ namespace ARMeilleure.Instructions
 
                 if (op.Size <= 2)
                 {
-                    de = EmitSatQ(context, emit(ne), op.Size, signedSrc: true, signedDst: true);
+                    de = EmitSignedSrcSatQ(context, emit(ne), op.Size, signedDst: true);
                 }
                 else /* if (op.Size == 3) */
                 {
@@ -1419,15 +1419,18 @@ namespace ARMeilleure.Instructions
                     {
                         Operand temp = add ? context.Add(ne, me) : context.Subtract(ne, me);
 
-                        de = EmitSatQ(context, temp, op.Size, signedSrc: true, signedDst: signed);
+                        de = EmitSignedSrcSatQ(context, temp, op.Size, signedDst: signed);
                     }
-                    else if (add) /* if (op.Size == 3) */
+                    else /* if (op.Size == 3) */
                     {
-                        de = EmitBinarySatQAdd(context, ne, me, signed);
+                        if (add)
+                        {
+                            de = signed ? EmitBinarySignedSatQAdd(context, ne, me) : EmitBinaryUnsignedSatQAdd(context, ne, me);
                         }
                         else /* if (sub) */
                         {
-                        de = EmitBinarySatQSub(context, ne, me, signed);
+                            de = signed ? EmitBinarySignedSatQSub(context, ne, me) : EmitBinaryUnsignedSatQSub(context, ne, me);
+                        }
                     }
 
                     res = EmitVectorInsert(context, res, de, index, op.Size);
@@ -1445,11 +1448,11 @@ namespace ARMeilleure.Instructions
                     {
                         Operand temp = context.Add(ne, me);
 
-                        de = EmitSatQ(context, temp, op.Size, signedSrc: true, signedDst: signed);
+                        de = EmitSignedSrcSatQ(context, temp, op.Size, signedDst: signed);
                     }
                     else /* if (op.Size == 3) */
                     {
-                        de = EmitBinarySatQAccumulate(context, ne, me, signed);
+                        de = signed ? EmitBinarySignedSatQAcc(context, ne, me) : EmitBinaryUnsignedSatQAcc(context, ne, me);
                     }
 
                     res = EmitVectorInsert(context, res, de, index, op.Size);
@@ -1475,7 +1478,7 @@ namespace ARMeilleure.Instructions
                         me = EmitVectorExtract(context, ((OpCodeSimdReg)op).Rm, index, op.Size, signed);
                     }
 
-                    Operand de = EmitSatQ(context, emit(ne, me), op.Size, true, signed);
+                    Operand de = EmitSignedSrcSatQ(context, emit(ne, me), op.Size, signedDst: signed);
 
                     res = EmitVectorInsert(context, res, de, index, op.Size);
                 }
@@ -1520,7 +1523,9 @@ namespace ARMeilleure.Instructions
             {
                 Operand ne = EmitVectorExtract(context, op.Rn, index, op.Size + 1, signedSrc);
 
-                Operand temp = EmitSatQ(context, ne, op.Size, signedSrc, signedDst);
+                Operand temp = signedSrc
+                    ? EmitSignedSrcSatQ(context, ne, op.Size, signedDst)
+                    : EmitUnsignedSrcSatQ(context, ne, op.Size, signedDst);
 
                 res = EmitVectorInsert(context, res, temp, part + index, op.Size);
             }
@@ -1528,74 +1533,248 @@ namespace ARMeilleure.Instructions
             context.Copy(d, res);
         }
 
-        // TSrc (16bit, 32bit, 64bit; signed, unsigned) > TDst (8bit, 16bit, 32bit; signed, unsigned).
-        public static Operand EmitSatQ(ArmEmitterContext context, Operand op, int sizeDst, bool signedSrc, bool signedDst)
+        // TSrc (16bit, 32bit, 64bit; signed) > TDst (8bit, 16bit, 32bit; signed, unsigned).
+        // long SignedSrcSignedDstSatQ(long op, int size); ulong SignedSrcUnsignedDstSatQ(long op, int size);
+        public static Operand EmitSignedSrcSatQ(ArmEmitterContext context, Operand op, int sizeDst, bool signedDst)
         {
-            if ((uint)sizeDst > 2u)
-            {
-                throw new ArgumentOutOfRangeException(nameof(sizeDst));
+            Debug.Assert(op.Type == OperandType.I64 && (uint)sizeDst <= 2u);
+
+            Operand lbl1 = Label();
+            Operand lblEnd = Label();
+
+            int eSize = 8 << sizeDst;
+
+            Operand maxT = signedDst ? Const((1L << (eSize - 1)) - 1L) : Const((1UL << eSize) - 1UL);
+            Operand minT = signedDst ? Const(-(1L << (eSize - 1))) : Const(0UL);
+
+            Operand res = context.Copy(context.AllocateLocal(OperandType.I64), op);
+
+            context.BranchIf(lbl1, res, maxT, Comparison.LessOrEqual);
+            context.Copy(res, maxT);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lbl1);
+            context.BranchIf(lblEnd, res, minT, Comparison.GreaterOrEqual);
+            context.Copy(res, minT);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lblEnd);
+
+            return res;
         }
 
-            MethodInfo info;
+        // TSrc (16bit, 32bit, 64bit; unsigned) > TDst (8bit, 16bit, 32bit; signed, unsigned).
+        // long UnsignedSrcSignedDstSatQ(ulong op, int size); ulong UnsignedSrcUnsignedDstSatQ(ulong op, int size);
+        public static Operand EmitUnsignedSrcSatQ(ArmEmitterContext context, Operand op, int sizeDst, bool signedDst)
+        {
+            Debug.Assert(op.Type == OperandType.I64 && (uint)sizeDst <= 2u);
 
-            if (signedSrc)
-            {
-                info = signedDst
-                    ? typeof(SoftFallback).GetMethod(nameof(SoftFallback.SignedSrcSignedDstSatQ))
-                    : typeof(SoftFallback).GetMethod(nameof(SoftFallback.SignedSrcUnsignedDstSatQ));
-            }
-            else
-            {
-                info = signedDst
-                    ? typeof(SoftFallback).GetMethod(nameof(SoftFallback.UnsignedSrcSignedDstSatQ))
-                    : typeof(SoftFallback).GetMethod(nameof(SoftFallback.UnsignedSrcUnsignedDstSatQ));
+            Operand lblEnd = Label();
+
+            int eSize = 8 << sizeDst;
+
+            Operand maxL = signedDst ? Const((1L << (eSize - 1)) - 1L) : Const((1UL << eSize) - 1UL);
+
+            Operand res = context.Copy(context.AllocateLocal(OperandType.I64), op);
+
+            context.BranchIf(lblEnd, res, maxL, Comparison.LessOrEqualUI);
+            context.Copy(res, maxL);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lblEnd);
+
+            return res;
         }
 
-            return context.Call(info, op, Const(sizeDst));
+        // long UnarySignedSatQAbsOrNeg(long op);
+        private static Operand EmitUnarySignedSatQAbsOrNeg(ArmEmitterContext context, Operand op)
+        {
+            Debug.Assert(op.Type == OperandType.I64);
+
+            Operand lblEnd = Label();
+
+            Operand minL = Const(long.MinValue);
+            Operand maxL = Const(long.MaxValue);
+
+            Operand res = context.Copy(context.AllocateLocal(OperandType.I64), op);
+
+            context.BranchIf(lblEnd, res, minL, Comparison.NotEqual);
+            context.Copy(res, maxL);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lblEnd);
+
+            return res;
         }
 
-        // TSrc (64bit) == TDst (64bit); signed.
-        public static Operand EmitUnarySignedSatQAbsOrNeg(ArmEmitterContext context, Operand op)
+        // long BinarySignedSatQAdd(long op1, long op2);
+        public static Operand EmitBinarySignedSatQAdd(ArmEmitterContext context, Operand op1, Operand op2)
         {
-            Debug.Assert(((OpCodeSimd)context.CurrOp).Size == 3, "Invalid element size.");
+            Debug.Assert(op1.Type == OperandType.I64 && op2.Type == OperandType.I64);
 
-            return context.Call(typeof(SoftFallback).GetMethod(nameof(SoftFallback.UnarySignedSatQAbsOrNeg)), op);
+            Operand lblEnd = Label();
+
+            Operand minL = Const(long.MinValue);
+            Operand maxL = Const(long.MaxValue);
+            Operand zero = Const(0L);
+
+            Operand res = context.Copy(context.AllocateLocal(OperandType.I64), context.Add(op1, op2));
+
+            Operand left = context.BitwiseNot(context.BitwiseExclusiveOr(op1, op2));
+            Operand right = context.BitwiseExclusiveOr(op1, res);
+            context.BranchIf(lblEnd, context.BitwiseAnd(left, right), zero, Comparison.GreaterOrEqual);
+
+            Operand isPositive = context.ICompareGreaterOrEqual(op1, zero);
+            context.Copy(res, context.ConditionalSelect(isPositive, maxL, minL));
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lblEnd);
+
+            return res;
         }
 
-        // TSrcs (64bit) == TDst (64bit); signed, unsigned.
-        public static Operand EmitBinarySatQAdd(ArmEmitterContext context, Operand op1, Operand op2, bool signed)
+        // ulong BinaryUnsignedSatQAdd(ulong op1, ulong op2);
+        public static Operand EmitBinaryUnsignedSatQAdd(ArmEmitterContext context, Operand op1, Operand op2)
         {
-            Debug.Assert(((OpCodeSimd)context.CurrOp).Size == 3, "Invalid element size.");
+            Debug.Assert(op1.Type == OperandType.I64 && op2.Type == OperandType.I64);
 
-            MethodInfo info = signed
-                ? typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinarySignedSatQAdd))
-                : typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinaryUnsignedSatQAdd));
+            Operand lblEnd = Label();
 
-            return context.Call(info, op1, op2);
+            Operand maxUL = Const(ulong.MaxValue);
+
+            Operand res = context.Copy(context.AllocateLocal(OperandType.I64), context.Add(op1, op2));
+
+            context.BranchIf(lblEnd, res, op1, Comparison.GreaterOrEqualUI);
+            context.Copy(res, maxUL);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lblEnd);
+
+            return res;
         }
 
-        // TSrcs (64bit) == TDst (64bit); signed, unsigned.
-        public static Operand EmitBinarySatQSub(ArmEmitterContext context, Operand op1, Operand op2, bool signed)
+        // long BinarySignedSatQSub(long op1, long op2);
+        public static Operand EmitBinarySignedSatQSub(ArmEmitterContext context, Operand op1, Operand op2)
         {
-            Debug.Assert(((OpCodeSimd)context.CurrOp).Size == 3, "Invalid element size.");
+            Debug.Assert(op1.Type == OperandType.I64 && op2.Type == OperandType.I64);
 
-            MethodInfo info = signed
-                ? typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinarySignedSatQSub))
-                : typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinaryUnsignedSatQSub));
+            Operand lblEnd = Label();
 
-            return context.Call(info, op1, op2);
+            Operand minL = Const(long.MinValue);
+            Operand maxL = Const(long.MaxValue);
+            Operand zero = Const(0L);
+
+            Operand res = context.Copy(context.AllocateLocal(OperandType.I64), context.Subtract(op1, op2));
+
+            Operand left = context.BitwiseExclusiveOr(op1, op2);
+            Operand right = context.BitwiseExclusiveOr(op1, res);
+            context.BranchIf(lblEnd, context.BitwiseAnd(left, right), zero, Comparison.GreaterOrEqual);
+
+            Operand isPositive = context.ICompareGreaterOrEqual(op1, zero);
+            context.Copy(res, context.ConditionalSelect(isPositive, maxL, minL));
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lblEnd);
+
+            return res;
         }
 
-        // TSrcs (64bit) == TDst (64bit); signed, unsigned.
-        public static Operand EmitBinarySatQAccumulate(ArmEmitterContext context, Operand op1, Operand op2, bool signed)
+        // ulong BinaryUnsignedSatQSub(ulong op1, ulong op2);
+        public static Operand EmitBinaryUnsignedSatQSub(ArmEmitterContext context, Operand op1, Operand op2)
         {
-            Debug.Assert(((OpCodeSimd)context.CurrOp).Size == 3, "Invalid element size.");
+            Debug.Assert(op1.Type == OperandType.I64 && op2.Type == OperandType.I64);
 
-            MethodInfo info = signed
-                ? typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinarySignedSatQAcc))
-                : typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinaryUnsignedSatQAcc));
+            Operand lblEnd = Label();
 
-            return context.Call(info, op1, op2);
+            Operand zero = Const(0L);
+
+            Operand res = context.Copy(context.AllocateLocal(OperandType.I64), context.Subtract(op1, op2));
+
+            context.BranchIf(lblEnd, op1, op2, Comparison.GreaterOrEqualUI);
+            context.Copy(res, zero);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lblEnd);
+
+            return res;
+        }
+
+        // long BinarySignedSatQAcc(ulong op1, long op2);
+        private static Operand EmitBinarySignedSatQAcc(ArmEmitterContext context, Operand op1, Operand op2)
+        {
+            Debug.Assert(op1.Type == OperandType.I64 && op2.Type == OperandType.I64);
+
+            Operand lbl1 = Label();
+            Operand lbl2 = Label();
+            Operand lblEnd = Label();
+
+            Operand maxL = Const(long.MaxValue);
+            Operand zero = Const(0L);
+
+            Operand res = context.Copy(context.AllocateLocal(OperandType.I64), context.Add(op1, op2));
+
+            context.BranchIf(lbl1, op1, maxL, Comparison.GreaterUI);
+            Operand notOp2AndRes = context.BitwiseAnd(context.BitwiseNot(op2), res);
+            context.BranchIf(lblEnd, notOp2AndRes, zero, Comparison.GreaterOrEqual);
+            context.Copy(res, maxL);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lbl1);
+            context.BranchIf(lbl2, op2, zero, Comparison.Less);
+            context.Copy(res, maxL);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lbl2);
+            context.BranchIf(lblEnd, res, maxL, Comparison.LessOrEqualUI);
+            context.Copy(res, maxL);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lblEnd);
+
+            return res;
+        }
+
+        // ulong BinaryUnsignedSatQAcc(long op1, ulong op2);
+        private static Operand EmitBinaryUnsignedSatQAcc(ArmEmitterContext context, Operand op1, Operand op2)
+        {
+            Debug.Assert(op1.Type == OperandType.I64 && op2.Type == OperandType.I64);
+
+            Operand lbl1 = Label();
+            Operand lblEnd = Label();
+
+            Operand maxUL = Const(ulong.MaxValue);
+            Operand maxL = Const(long.MaxValue);
+            Operand zero = Const(0L);
+
+            Operand res = context.Copy(context.AllocateLocal(OperandType.I64), context.Add(op1, op2));
+
+            context.BranchIf(lbl1, op1, zero, Comparison.Less);
+            context.BranchIf(lblEnd, res, op1, Comparison.GreaterOrEqualUI);
+            context.Copy(res, maxUL);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lbl1);
+            context.BranchIf(lblEnd, op2, maxL, Comparison.GreaterUI);
+            context.BranchIf(lblEnd, res, zero, Comparison.GreaterOrEqual);
+            context.Copy(res, zero);
+            context.Call(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
+            context.Branch(lblEnd);
+
+            context.MarkLabel(lblEnd);
+
+            return res;
         }
 
         public static Operand EmitFloatAbs(ArmEmitterContext context, Operand value, bool single, bool vector)

ARMeilleure/Instructions/InstEmitSimdHelper32.cs

@@ -219,6 +219,25 @@ namespace ARMeilleure.Instructions
 
         // Integer
 
+        public static void EmitVectorUnaryAccumulateOpI32(ArmEmitterContext context, Func1I emit, bool signed)
+        {
+            OpCode32Simd op = (OpCode32Simd)context.CurrOp;
+
+            Operand res = GetVecA32(op.Qd);
+
+            int elems = op.GetBytesCount() >> op.Size;
+
+            for (int index = 0; index < elems; index++)
+            {
+                Operand de = EmitVectorExtract32(context, op.Qd, op.Id + index, op.Size, signed);
+                Operand me = EmitVectorExtract32(context, op.Qm, op.Im + index, op.Size, signed);
+
+                res = EmitVectorInsert(context, res, context.Add(de, emit(me)), op.Id + index, op.Size);
+            }
+
+            context.Copy(GetVecA32(op.Qd), res);
+        }
+
         public static void EmitVectorUnaryOpI32(ArmEmitterContext context, Func1I emit, bool signed)
         {
             OpCode32Simd op = (OpCode32Simd)context.CurrOp;
@@ -385,6 +404,18 @@ namespace ARMeilleure.Instructions
             EmitVectorUnaryOpI32(context, emit, true);
         }
 
+        public static void EmitVectorUnaryOpSx32(ArmEmitterContext context, Func1I emit, bool accumulate)
+        {
+            if (accumulate)
+            {
+                EmitVectorUnaryAccumulateOpI32(context, emit, true);
+            }
+            else
+            {
+                EmitVectorUnaryOpI32(context, emit, true);
+            }
+        }
+
         public static void EmitVectorBinaryOpSx32(ArmEmitterContext context, Func2I emit)
         {
             EmitVectorBinaryOpI32(context, emit, true);
@@ -400,6 +431,18 @@ namespace ARMeilleure.Instructions
             EmitVectorUnaryOpI32(context, emit, false);
         }
 
+        public static void EmitVectorUnaryOpZx32(ArmEmitterContext context, Func1I emit, bool accumulate)
+        {
+            if (accumulate)
+            {
+                EmitVectorUnaryAccumulateOpI32(context, emit, false);
+            }
+            else
+            {
+                EmitVectorUnaryOpI32(context, emit, false);
+            }
+        }
+
         public static void EmitVectorBinaryOpZx32(ArmEmitterContext context, Func2I emit)
         {
             EmitVectorBinaryOpI32(context, emit, false);
@@ -592,6 +635,34 @@ namespace ARMeilleure.Instructions
             context.Copy(GetVecA32(op.Qd), res);
         }
 
+        public static void EmitVectorPairwiseLongOpI32(ArmEmitterContext context, Func2I emit, bool signed)
+        {
+            OpCode32Simd op = (OpCode32Simd)context.CurrOp;
+
+            int elems = (op.Q ? 16 : 8) >> op.Size;
+            int pairs = elems >> 1;
+            int id = (op.Vd & 1) * pairs;
+
+            Operand res = GetVecA32(op.Qd);
+
+            for (int index = 0; index < pairs; index++)
+            {
+                int pairIndex = index << 1;
+                Operand m1 = EmitVectorExtract32(context, op.Qm, op.Im + pairIndex, op.Size, signed);
+                Operand m2 = EmitVectorExtract32(context, op.Qm, op.Im + pairIndex + 1, op.Size, signed);
+
+                if (op.Size == 2)
+                {
+                    m1 = signed ? context.SignExtend32(OperandType.I64, m1) : context.ZeroExtend32(OperandType.I64, m1);
+                    m2 = signed ? context.SignExtend32(OperandType.I64, m2) : context.ZeroExtend32(OperandType.I64, m2);
+                }
+
+                res = EmitVectorInsert(context, res, emit(m1, m2), id + index, op.Size + 1);
+            }
+
+            context.Copy(GetVecA32(op.Qd), res);
+        }
+
         // Narrow
 
         public static void EmitVectorUnaryNarrowOp32(ArmEmitterContext context, Func1I emit, bool signed = false)

ARMeilleure/Instructions/InstEmitSimdShift.cs

@@ -1004,7 +1004,7 @@ namespace ARMeilleure.Instructions
                     e = EmitShrImm64(context, e, signedSrc, roundConst, shift); // shift <= 32
                 }
 
-                e = EmitSatQ(context, e, op.Size, signedSrc, signedDst);
+                e = signedSrc ? EmitSignedSrcSatQ(context, e, op.Size, signedDst) : EmitUnsignedSrcSatQ(context, e, op.Size, signedDst);
 
                 res = EmitVectorInsert(context, res, e, part + index, op.Size);
             }

ARMeilleure/Instructions/InstEmitSimdShift32.cs

@@ -33,64 +33,24 @@ namespace ARMeilleure.Instructions
             EmitShrImmSaturatingNarrowOp(context, op.U ? ShrImmSaturatingNarrowFlags.VectorZxZx : ShrImmSaturatingNarrowFlags.VectorSxSx);
         }
 
+        public static void Vqshrun(ArmEmitterContext context)
+        {
+            EmitShrImmSaturatingNarrowOp(context, ShrImmSaturatingNarrowFlags.VectorSxZx);
+        }
+
         public static void Vrshr(ArmEmitterContext context)
         {
-            OpCode32SimdShImm op = (OpCode32SimdShImm)context.CurrOp;
-            int shift = GetImmShr(op);
-            long roundConst = 1L << (shift - 1);
+            EmitRoundShrImmOp(context, accumulate: false);
+        }
 
-            if (op.U)
+        public static void Vrshrn(ArmEmitterContext context)
         {
-                if (op.Size < 2)
-                {
-                    EmitVectorUnaryOpZx32(context, (op1) =>
-                    {
-                        op1 = context.Add(op1, Const(op1.Type, roundConst));
+            EmitRoundShrImmNarrowOp(context, signed: false);
+        }
 
-                        return context.ShiftRightUI(op1, Const(shift));
-                    });
-                }
-                else if (op.Size == 2)
+        public static void Vrsra(ArmEmitterContext context)
         {
-                    EmitVectorUnaryOpZx32(context, (op1) =>
-                    {
-                        op1 = context.ZeroExtend32(OperandType.I64, op1);
-                        op1 = context.Add(op1, Const(op1.Type, roundConst));
-
-                        return context.ConvertI64ToI32(context.ShiftRightUI(op1, Const(shift)));
-                    });
-                }
-                else /* if (op.Size == 3) */
-                {
-                    EmitVectorUnaryOpZx32(context, (op1) => EmitShrImm64(context, op1, signed: false, roundConst, shift));
-                }
-            }
-            else
-            {
-                if (op.Size < 2)
-                {
-                    EmitVectorUnaryOpSx32(context, (op1) =>
-                    {
-                        op1 = context.Add(op1, Const(op1.Type, roundConst));
-
-                        return context.ShiftRightSI(op1, Const(shift));
-                    });
-                }
-                else if (op.Size == 2)
-                {
-                    EmitVectorUnaryOpSx32(context, (op1) =>
-                    {
-                        op1 = context.SignExtend32(OperandType.I64, op1);
-                        op1 = context.Add(op1, Const(op1.Type, roundConst));
-
-                        return context.ConvertI64ToI32(context.ShiftRightSI(op1, Const(shift)));
-                    });
-                }
-                else /* if (op.Size == 3) */
-                {
-                    EmitVectorUnaryOpZx32(context, (op1) => EmitShrImm64(context, op1, signed: true, roundConst, shift));
-                }
-            }
+            EmitRoundShrImmOp(context, accumulate: true);
         }
 
         public static void Vshl(ArmEmitterContext context)
@@ -191,6 +151,89 @@ namespace ARMeilleure.Instructions
             }
         }
 
+        public static void EmitRoundShrImmOp(ArmEmitterContext context, bool accumulate)
+        {
+            OpCode32SimdShImm op = (OpCode32SimdShImm)context.CurrOp;
+            int shift = GetImmShr(op);
+            long roundConst = 1L << (shift - 1);
+
+            if (op.U)
+            {
+                if (op.Size < 2)
+                {
+                    EmitVectorUnaryOpZx32(context, (op1) =>
+                    {
+                        op1 = context.Add(op1, Const(op1.Type, roundConst));
+
+                        return context.ShiftRightUI(op1, Const(shift));
+                    }, accumulate);
+                }
+                else if (op.Size == 2)
+                {
+                    EmitVectorUnaryOpZx32(context, (op1) =>
+                    {
+                        op1 = context.ZeroExtend32(OperandType.I64, op1);
+                        op1 = context.Add(op1, Const(op1.Type, roundConst));
+
+                        return context.ConvertI64ToI32(context.ShiftRightUI(op1, Const(shift)));
+                    }, accumulate);
+                }
+                else /* if (op.Size == 3) */
+                {
+                    EmitVectorUnaryOpZx32(context, (op1) => EmitShrImm64(context, op1, signed: false, roundConst, shift), accumulate);
+                }
+            }
+            else
+            {
+                if (op.Size < 2)
+                {
+                    EmitVectorUnaryOpSx32(context, (op1) =>
+                    {
+                        op1 = context.Add(op1, Const(op1.Type, roundConst));
+
+                        return context.ShiftRightSI(op1, Const(shift));
+                    }, accumulate);
+                }
+                else if (op.Size == 2)
+                {
+                    EmitVectorUnaryOpSx32(context, (op1) =>
+                    {
+                        op1 = context.SignExtend32(OperandType.I64, op1);
+                        op1 = context.Add(op1, Const(op1.Type, roundConst));
+
+                        return context.ConvertI64ToI32(context.ShiftRightSI(op1, Const(shift)));
+                    }, accumulate);
+                }
+                else /* if (op.Size == 3) */
+                {
+                    EmitVectorUnaryOpZx32(context, (op1) => EmitShrImm64(context, op1, signed: true, roundConst, shift), accumulate);
+                }
+            }
+        }
+
+        private static void EmitRoundShrImmNarrowOp(ArmEmitterContext context, bool signed)
+        {
+            OpCode32SimdShImm op = (OpCode32SimdShImm)context.CurrOp;
+
+            int shift = GetImmShr(op);
+            long roundConst = 1L << (shift - 1);
+
+            EmitVectorUnaryNarrowOp32(context, (op1) =>
+            {
+                if (op.Size <= 1)
+                {
+                    op1 = context.Add(op1, Const(op1.Type, roundConst));
+                    op1 = signed ? context.ShiftRightSI(op1, Const(shift)) : context.ShiftRightUI(op1, Const(shift));
+                }
+                else /* if (op.Size == 2 && round) */
+                {
+                    op1 = EmitShrImm64(context, op1, signed, roundConst, shift); // shift <= 32
+                }
+
+                return op1;
+            }, signed);
+        }
+
         private static Operand EmitShlRegOp(ArmEmitterContext context, Operand op, Operand shiftLsB, int size, bool unsigned)
         {
             if (shiftLsB.Type == OperandType.I64)
@@ -289,7 +332,7 @@ namespace ARMeilleure.Instructions
                     op1 = EmitShrImm64(context, op1, signedSrc, roundConst, shift); // shift <= 32
                 }
 
-                return EmitSatQ(context, op1, 8 << op.Size, signedDst);
+                return EmitSatQ(context, op1, 8 << op.Size, signedSrc, signedDst);
             }, signedSrc);
         }
 
@@ -313,15 +356,20 @@ namespace ARMeilleure.Instructions
             return context.Call(info, value, Const(roundConst), Const(shift));
         }
 
-        private static Operand EmitSatQ(ArmEmitterContext context, Operand value, int eSize, bool signed)
+        private static Operand EmitSatQ(ArmEmitterContext context, Operand value, int eSize, bool signedSrc, bool signedDst)
         {
             Debug.Assert(eSize <= 32);
 
-            long intMin = signed ? -(1L << (eSize - 1)) : 0;
-            long intMax = signed ? (1L << (eSize - 1)) - 1 : (1L << eSize) - 1;
+            long intMin = signedDst ? -(1L << (eSize - 1)) : 0;
+            long intMax = signedDst ? (1L << (eSize - 1)) - 1 : (1L << eSize) - 1;
 
-            Operand gt = context.ICompareGreater(value, Const(value.Type, intMax));
-            Operand lt = context.ICompareLess(value, Const(value.Type, intMin));
+            Operand gt = signedSrc
+                ? context.ICompareGreater(value, Const(value.Type, intMax))
+                : context.ICompareGreaterUI(value, Const(value.Type, intMax));
+
+            Operand lt = signedSrc
+                ? context.ICompareLess(value, Const(value.Type, intMin))
+                : context.ICompareLessUI(value, Const(value.Type, intMin));
 
             value = context.ConditionalSelect(gt, Const(value.Type, intMax), value);
             value = context.ConditionalSelect(lt, Const(value.Type, intMin), value);

ARMeilleure/Instructions/InstEmitSystem32.cs

@@ -16,18 +16,13 @@ namespace ARMeilleure.Instructions
         {
             OpCode32System op = (OpCode32System)context.CurrOp;
 
-            if (op.Coproc != 15)
+            if (op.Coproc != 15 || op.Opc1 != 0)
             {
                 InstEmit.Und(context);
 
                 return;
             }
 
-            if (op.Opc1 != 0)
-            {
-                throw new NotImplementedException($"Unknown MRC Opc1 0x{op.Opc1:X16} at 0x{op.Address:X16}.");
-            }
-
             MethodInfo info;
 
             switch (op.CRn)
@@ -35,7 +30,7 @@ namespace ARMeilleure.Instructions
                 case 13: // Process and Thread Info.
                     if (op.CRm != 0)
                     {
-                        throw new NotImplementedException($"Unknown MRC CRm 0x{op.CRm:X16} at 0x{op.Address:X16}.");
+                        throw new NotImplementedException($"Unknown MRC CRm 0x{op.CRm:X} at 0x{op.Address:X} (0x{op.RawOpCode:X}).");
                     }
 
                     switch (op.Opc2)
@@ -44,7 +39,7 @@ namespace ARMeilleure.Instructions
                             info = typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetTpidrEl032)); break;
 
                         default:
-                            throw new NotImplementedException($"Unknown MRC Opc2 0x{op.Opc2:X16} at 0x{op.Address:X16}.");
+                            throw new NotImplementedException($"Unknown MRC Opc2 0x{op.Opc2:X} at 0x{op.Address:X} (0x{op.RawOpCode:X}).");
                     }
 
                     break;
@@ -59,11 +54,11 @@ namespace ARMeilleure.Instructions
                                     return; // No-op.
 
                                 default:
-                                    throw new NotImplementedException($"Unknown MRC Opc2 0x{op.Opc2:X16} at 0x{op.Address:X16}.");
+                                    throw new NotImplementedException($"Unknown MRC Opc2 0x{op.Opc2:X16} at 0x{op.Address:X16} (0x{op.RawOpCode:X}).");
                             }
 
                         default:
-                            throw new NotImplementedException($"Unknown MRC CRm 0x{op.CRm:X16} at 0x{op.Address:X16}.");
+                            throw new NotImplementedException($"Unknown MRC CRm 0x{op.CRm:X16} at 0x{op.Address:X16} (0x{op.RawOpCode:X}).");
                     }
 
                 default:
@@ -77,18 +72,13 @@ namespace ARMeilleure.Instructions
         {
             OpCode32System op = (OpCode32System)context.CurrOp;
 
-            if (op.Coproc != 15)
+            if (op.Coproc != 15 || op.Opc1 != 0)
             {
                 InstEmit.Und(context);
 
                 return;
             }
 
-            if (op.Opc1 != 0)
-            {
-                throw new NotImplementedException($"Unknown MRC Opc1 0x{op.Opc1:X16} at 0x{op.Address:X16}.");
-            }
-
             MethodInfo info;
 
             switch (op.CRn)
@@ -96,7 +86,7 @@ namespace ARMeilleure.Instructions
                 case 13: // Process and Thread Info.
                     if (op.CRm != 0)
                     {
-                        throw new NotImplementedException($"Unknown MRC CRm 0x{op.CRm:X16} at 0x{op.Address:X16}.");
+                        throw new NotImplementedException($"Unknown MRC CRm 0x{op.CRm:X} at 0x{op.Address:X} (0x{op.RawOpCode:X}).");
                     }
 
                     switch (op.Opc2)
@@ -108,13 +98,13 @@ namespace ARMeilleure.Instructions
                             info = typeof(NativeInterface).GetMethod(nameof(NativeInterface.GetTpidr32)); break;
 
                         default:
-                            throw new NotImplementedException($"Unknown MRC Opc2 0x{op.Opc2:X16} at 0x{op.Address:X16}.");
+                            throw new NotImplementedException($"Unknown MRC Opc2 0x{op.Opc2:X} at 0x{op.Address:X} (0x{op.RawOpCode:X}).");
                     }
 
                     break;
 
                 default:
-                    throw new NotImplementedException($"Unknown MRC 0x{op.RawOpCode:X8} at 0x{op.Address:X16}.");
+                    throw new NotImplementedException($"Unknown MRC 0x{op.RawOpCode:X} at 0x{op.Address:X}.");
             }
 
             if (op.Rt == RegisterAlias.Aarch32Pc)
@@ -154,13 +144,13 @@ namespace ARMeilleure.Instructions
                             info = typeof(NativeInterface).GetMethod(nameof(NativeInterface.GetCntpctEl0)); break;
 
                         default:
-                            throw new NotImplementedException($"Unknown MRRC Opc1 0x{opc:X16} at 0x{op.Address:X16}.");
+                            throw new NotImplementedException($"Unknown MRRC Opc1 0x{opc:X} at 0x{op.Address:X} (0x{op.RawOpCode:X}).");
                     }
 
                     break;
 
                 default:
-                    throw new NotImplementedException($"Unknown MRRC 0x{op.RawOpCode:X8} at 0x{op.Address:X16}.");
+                    throw new NotImplementedException($"Unknown MRRC 0x{op.RawOpCode:X} at 0x{op.Address:X}.");
             }
 
             Operand result = context.Call(info);
@@ -169,6 +159,31 @@ namespace ARMeilleure.Instructions
             SetIntA32(context, op.CRn, context.ConvertI64ToI32(context.ShiftRightUI(result, Const(32))));
         }
 
+        public static void Mrs(ArmEmitterContext context)
+        {
+            OpCode32Mrs op = (OpCode32Mrs)context.CurrOp;
+
+            if (op.R)
+            {
+                throw new NotImplementedException("SPSR");
+            }
+            else
+            {
+                Operand vSh = context.ShiftLeft(GetFlag(PState.VFlag), Const((int)PState.VFlag));
+                Operand cSh = context.ShiftLeft(GetFlag(PState.CFlag), Const((int)PState.CFlag));
+                Operand zSh = context.ShiftLeft(GetFlag(PState.ZFlag), Const((int)PState.ZFlag));
+                Operand nSh = context.ShiftLeft(GetFlag(PState.NFlag), Const((int)PState.NFlag));
+                Operand qSh = context.ShiftLeft(GetFlag(PState.QFlag), Const((int)PState.QFlag));
+
+                Operand spsr = context.BitwiseOr(context.BitwiseOr(nSh, zSh), context.BitwiseOr(cSh, vSh));
+                spsr = context.BitwiseOr(spsr, qSh);
+
+                // TODO: Remaining flags.
+
+                SetIntA32(context, op.Rd, spsr);
+            }
+        }
+
         public static void Msr(ArmEmitterContext context)
         {
             OpCode32MsrReg op = (OpCode32MsrReg)context.CurrOp;
@@ -240,7 +255,7 @@ namespace ARMeilleure.Instructions
                 case 0b1000: // FPEXC
                     throw new NotImplementedException("Supervisor Only");
                 default:
-                    throw new NotImplementedException($"Unknown VMRS 0x{op.RawOpCode:X8} at 0x{op.Address:X16}.");
+                    throw new NotImplementedException($"Unknown VMRS 0x{op.RawOpCode:X} at 0x{op.Address:X}.");
             }
         }
 
@@ -263,7 +278,7 @@ namespace ARMeilleure.Instructions
                 case 0b1000: // FPEXC
                     throw new NotImplementedException("Supervisor Only");
                 default:
-                    throw new NotImplementedException($"Unknown VMSR 0x{op.RawOpCode:X8} at 0x{op.Address:X16}.");
+                    throw new NotImplementedException($"Unknown VMSR 0x{op.RawOpCode:X} at 0x{op.Address:X}.");
             }
         }
 

ARMeilleure/Instructions/InstName.cs

@@ -545,6 +545,8 @@ namespace ARMeilleure.Instructions
         Strexh,
         Strh,
         Sxtb16,
+        Tbb,
+        Tbh,
         Teq,
         Trap,
         Tst,
@@ -601,6 +603,7 @@ namespace ARMeilleure.Instructions
         Vmin,
         Vminnm,
         Vmla,
+        Vmlal,
         Vmls,
         Vmlsl,
         Vmov,
@@ -618,15 +621,24 @@ namespace ARMeilleure.Instructions
         Vorn,
         Vorr,
         Vpadd,
+        Vpaddl,
         Vpmax,
         Vpmin,
+        Vqadd,
+        Vqdmulh,
+        Vqmovn,
+        Vqmovun,
         Vqrshrn,
         Vqrshrun,
         Vqshrn,
+        Vqshrun,
+        Vqsub,
         Vrev,
+        Vrhadd,
         Vrint,
         Vrintx,
         Vrshr,
+        Vrshrn,
         Vsel,
         Vshl,
         Vshll,
@@ -643,8 +655,10 @@ namespace ARMeilleure.Instructions
         Vrecps,
         Vrsqrte,
         Vrsqrts,
+        Vrsra,
         Vsra,
         Vsub,
+        Vsubl,
         Vsubw,
         Vtbl,
         Vtrn,

ARMeilleure/Instructions/NativeInterface.cs

@@ -2,8 +2,6 @@ using ARMeilleure.Memory;
 using ARMeilleure.State;
 using ARMeilleure.Translation;
 using System;
-using System.Diagnostics;
-using System.Runtime.InteropServices;
 
 namespace ARMeilleure.Instructions
 {

ARMeilleure/Instructions/SoftFallback.cs

@@ -91,7 +91,7 @@ namespace ARMeilleure.Instructions
                 }
                 else /* if (eSize != 64) */
                 {
-                    return SignedSrcSignedDstSatQ(value << shiftLsB, size);
+                    return SignedSrcSignedDstSatQ(value << shiftLsB, size); // InstEmitSimdHelper.EmitSignedSrcSatQ(signedDst: true).
                 }
             }
             else /* if (shiftLsB == 0) */
@@ -135,7 +135,7 @@ namespace ARMeilleure.Instructions
                 }
                 else /* if (eSize != 64) */
                 {
-                    return UnsignedSrcUnsignedDstSatQ(value << shiftLsB, size);
+                    return UnsignedSrcUnsignedDstSatQ(value << shiftLsB, size); // InstEmitSimdHelper.EmitUnsignedSrcSatQ(signedDst: false).
                 }
             }
             else /* if (shiftLsB == 0) */
@@ -509,7 +509,7 @@ namespace ARMeilleure.Instructions
 #endregion
 
 #region "Saturating"
-        public static long SignedSrcSignedDstSatQ(long op, int size)
+        private static long SignedSrcSignedDstSatQ(long op, int size)
         {
             ExecutionContext context = NativeInterface.GetContext();
 
@@ -536,54 +536,7 @@ namespace ARMeilleure.Instructions
             }
         }
 
-        public static ulong SignedSrcUnsignedDstSatQ(long op, int size)
-        {
-            ExecutionContext context = NativeInterface.GetContext();
-
-            int eSize = 8 << size;
-
-            ulong tMaxValue = (1UL << eSize) - 1UL;
-            ulong tMinValue =  0UL;
-
-            if (op > (long)tMaxValue)
-            {
-                context.Fpsr |= FPSR.Qc;
-
-                return tMaxValue;
-            }
-            else if (op < (long)tMinValue)
-            {
-                context.Fpsr |= FPSR.Qc;
-
-                return tMinValue;
-            }
-            else
-            {
-                return (ulong)op;
-            }
-        }
-
-        public static long UnsignedSrcSignedDstSatQ(ulong op, int size)
-        {
-            ExecutionContext context = NativeInterface.GetContext();
-
-            int eSize = 8 << size;
-
-            long tMaxValue = (1L << (eSize - 1)) - 1L;
-
-            if (op > (ulong)tMaxValue)
-            {
-                context.Fpsr |= FPSR.Qc;
-
-                return tMaxValue;
-            }
-            else
-            {
-                return (long)op;
-            }
-        }
-
-        public static ulong UnsignedSrcUnsignedDstSatQ(ulong op, int size)
+        private static ulong UnsignedSrcUnsignedDstSatQ(ulong op, int size)
         {
             ExecutionContext context = NativeInterface.GetContext();
 
@@ -602,208 +555,6 @@ namespace ARMeilleure.Instructions
                 return op;
             }
         }
-
-        public static long UnarySignedSatQAbsOrNeg(long op)
-        {
-            ExecutionContext context = NativeInterface.GetContext();
-
-            if (op == long.MinValue)
-            {
-                context.Fpsr |= FPSR.Qc;
-
-                return long.MaxValue;
-            }
-            else
-            {
-                return op;
-            }
-        }
-
-        public static long BinarySignedSatQAdd(long op1, long op2)
-        {
-            ExecutionContext context = NativeInterface.GetContext();
-
-            long add = op1 + op2;
-
-            if ((~(op1 ^ op2) & (op1 ^ add)) < 0L)
-            {
-                context.Fpsr |= FPSR.Qc;
-
-                if (op1 < 0L)
-                {
-                    return long.MinValue;
-                }
-                else
-                {
-                    return long.MaxValue;
-                }
-            }
-            else
-            {
-                return add;
-            }
-        }
-
-        public static ulong BinaryUnsignedSatQAdd(ulong op1, ulong op2)
-        {
-            ExecutionContext context = NativeInterface.GetContext();
-
-            ulong add = op1 + op2;
-
-            if ((add < op1) && (add < op2))
-            {
-                context.Fpsr |= FPSR.Qc;
-
-                return ulong.MaxValue;
-            }
-            else
-            {
-                return add;
-            }
-        }
-
-        public static long BinarySignedSatQSub(long op1, long op2)
-        {
-            ExecutionContext context = NativeInterface.GetContext();
-
-            long sub = op1 - op2;
-
-            if (((op1 ^ op2) & (op1 ^ sub)) < 0L)
-            {
-                context.Fpsr |= FPSR.Qc;
-
-                if (op1 < 0L)
-                {
-                    return long.MinValue;
-                }
-                else
-                {
-                    return long.MaxValue;
-                }
-            }
-            else
-            {
-                return sub;
-            }
-        }
-
-        public static ulong BinaryUnsignedSatQSub(ulong op1, ulong op2)
-        {
-            ExecutionContext context = NativeInterface.GetContext();
-
-            ulong sub = op1 - op2;
-
-            if (op1 < op2)
-            {
-                context.Fpsr |= FPSR.Qc;
-
-                return ulong.MinValue;
-            }
-            else
-            {
-                return sub;
-            }
-        }
-
-        public static long BinarySignedSatQAcc(ulong op1, long op2)
-        {
-            ExecutionContext context = NativeInterface.GetContext();
-
-            if (op1 <= (ulong)long.MaxValue)
-            {
-                // op1 from ulong.MinValue to (ulong)long.MaxValue
-                // op2 from long.MinValue to long.MaxValue
-
-                long add = (long)op1 + op2;
-
-                if ((~op2 & add) < 0L)
-                {
-                    context.Fpsr |= FPSR.Qc;
-
-                    return long.MaxValue;
-                }
-                else
-                {
-                    return add;
-                }
-            }
-            else if (op2 >= 0L)
-            {
-                // op1 from (ulong)long.MaxValue + 1UL to ulong.MaxValue
-                // op2 from (long)ulong.MinValue to long.MaxValue
-
-                context.Fpsr |= FPSR.Qc;
-
-                return long.MaxValue;
-            }
-            else
-            {
-                // op1 from (ulong)long.MaxValue + 1UL to ulong.MaxValue
-                // op2 from long.MinValue to (long)ulong.MinValue - 1L
-
-                ulong add = op1 + (ulong)op2;
-
-                if (add > (ulong)long.MaxValue)
-                {
-                    context.Fpsr |= FPSR.Qc;
-
-                    return long.MaxValue;
-                }
-                else
-                {
-                    return (long)add;
-                }
-            }
-        }
-
-        public static ulong BinaryUnsignedSatQAcc(long op1, ulong op2)
-        {
-            ExecutionContext context = NativeInterface.GetContext();
-
-            if (op1 >= 0L)
-            {
-                // op1 from (long)ulong.MinValue to long.MaxValue
-                // op2 from ulong.MinValue to ulong.MaxValue
-
-                ulong add = (ulong)op1 + op2;
-
-                if ((add < (ulong)op1) && (add < op2))
-                {
-                    context.Fpsr |= FPSR.Qc;
-
-                    return ulong.MaxValue;
-                }
-                else
-                {
-                    return add;
-                }
-            }
-            else if (op2 > (ulong)long.MaxValue)
-            {
-                // op1 from long.MinValue to (long)ulong.MinValue - 1L
-                // op2 from (ulong)long.MaxValue + 1UL to ulong.MaxValue
-
-                return (ulong)op1 + op2;
-            }
-            else
-            {
-                // op1 from long.MinValue to (long)ulong.MinValue - 1L
-                // op2 from ulong.MinValue to (ulong)long.MaxValue
-
-                long add = op1 + (long)op2;
-
-                if (add < (long)ulong.MinValue)
-                {
-                    context.Fpsr |= FPSR.Qc;
-
-                    return ulong.MinValue;
-                }
-                else
-                {
-                    return (ulong)add;
-                }
-            }
-        }
 #endregion
 
 #region "Count"
@@ -1129,7 +880,7 @@ namespace ARMeilleure.Instructions
             return Sha256Hash(hash_abcd, hash_efgh, wk, part1: true);
         }
 
-        public static V128 HashUpper(V128 hash_efgh, V128 hash_abcd, V128 wk)
+        public static V128 HashUpper(V128 hash_abcd, V128 hash_efgh, V128 wk)
         {
             return Sha256Hash(hash_abcd, hash_efgh, wk, part1: false);
         }

ARMeilleure/IntermediateRepresentation/Intrinsic.cs

@@ -71,6 +71,7 @@ namespace ARMeilleure.IntermediateRepresentation
         X86Paddd,
         X86Paddq,
         X86Paddw,
+        X86Palignr,
         X86Pand,
         X86Pandn,
         X86Pavgb,
@@ -140,6 +141,9 @@ namespace ARMeilleure.IntermediateRepresentation
         X86Roundss,
         X86Rsqrtps,
         X86Rsqrtss,
+        X86Sha256Msg1,
+        X86Sha256Msg2,
+        X86Sha256Rnds2,
         X86Shufpd,
         X86Shufps,
         X86Sqrtpd,

ARMeilleure/Optimizations.cs

@@ -21,6 +21,7 @@ namespace ARMeilleure
         public static bool UseFmaIfAvailable       { get; set; } = true;
         public static bool UseAesniIfAvailable     { get; set; } = true;
         public static bool UsePclmulqdqIfAvailable { get; set; } = true;
+        public static bool UseShaIfAvailable       { get; set; } = true;
 
         public static bool ForceLegacySse
         {
@@ -40,5 +41,6 @@ namespace ARMeilleure
         internal static bool UseFma       => UseFmaIfAvailable       && HardwareCapabilities.SupportsFma;
         internal static bool UseAesni     => UseAesniIfAvailable     && HardwareCapabilities.SupportsAesni;
         internal static bool UsePclmulqdq => UsePclmulqdqIfAvailable && HardwareCapabilities.SupportsPclmulqdq;
+        internal static bool UseSha       => UseShaIfAvailable       && HardwareCapabilities.SupportsSha;
     }
 }

ARMeilleure/Signal/NativeSignalHandler.cs

@@ -197,12 +197,29 @@ namespace ARMeilleure.Signal
                 // Only call tracking if in range.
                 context.BranchIfFalse(nextLabel, inRange, BasicBlockFrequency.Cold);
 
-                context.Copy(inRegionLocal, Const(1));
                 Operand offset = context.BitwiseAnd(context.Subtract(faultAddress, rangeAddress), Const(~PageMask));
 
                 // Call the tracking action, with the pointer's relative offset to the base address.
                 Operand trackingActionPtr = context.Load(OperandType.I64, Const((ulong)signalStructPtr + rangeBaseOffset + 20));
-                context.Call(trackingActionPtr, OperandType.I32, offset, Const(PageSize), isWrite, Const(0));
+
+                context.Copy(inRegionLocal, Const(0));
+
+                Operand skipActionLabel = Label();
+
+                // Tracking action should be non-null to call it, otherwise assume false return.
+                context.BranchIfFalse(skipActionLabel, trackingActionPtr);
+                Operand result = context.Call(trackingActionPtr, OperandType.I32, offset, Const(PageSize), isWrite, Const(0));
+                context.Copy(inRegionLocal, result);
+
+                context.MarkLabel(skipActionLabel);
+
+                // If the tracking action returns false or does not exist, it might be an invalid access due to a partial overlap on Windows.
+                if (OperatingSystem.IsWindows())
+                {
+                    context.BranchIfTrue(endLabel, inRegionLocal);
+
+                    context.Copy(inRegionLocal, WindowsPartialUnmapHandler.EmitRetryFromAccessViolation(context));
+                }
 
                 context.Branch(endLabel);
 

ARMeilleure/Signal/TestMethods.cs

@@ -0,0 +1,84 @@
+using ARMeilleure.IntermediateRepresentation;
+using ARMeilleure.Translation;
+using System;
+
+using static ARMeilleure.IntermediateRepresentation.Operand.Factory;
+
+namespace ARMeilleure.Signal
+{
+    public struct NativeWriteLoopState
+    {
+        public int Running;
+        public int Error;
+    }
+
+    public static class TestMethods
+    {
+        public delegate bool DebugPartialUnmap();
+        public delegate int DebugThreadLocalMapGetOrReserve(int threadId, int initialState);
+        public delegate void DebugNativeWriteLoop(IntPtr nativeWriteLoopPtr, IntPtr writePtr);
+
+        public static DebugPartialUnmap GenerateDebugPartialUnmap()
+        {
+            EmitterContext context = new EmitterContext();
+
+            var result = WindowsPartialUnmapHandler.EmitRetryFromAccessViolation(context);
+
+            context.Return(result);
+
+            // Compile and return the function.
+
+            ControlFlowGraph cfg = context.GetControlFlowGraph();
+
+            OperandType[] argTypes = new OperandType[] { OperandType.I64 };
+
+            return Compiler.Compile(cfg, argTypes, OperandType.I32, CompilerOptions.HighCq).Map<DebugPartialUnmap>();
+        }
+
+        public static DebugThreadLocalMapGetOrReserve GenerateDebugThreadLocalMapGetOrReserve(IntPtr structPtr)
+        {
+            EmitterContext context = new EmitterContext();
+
+            var result = WindowsPartialUnmapHandler.EmitThreadLocalMapIntGetOrReserve(context, structPtr, context.LoadArgument(OperandType.I32, 0), context.LoadArgument(OperandType.I32, 1));
+
+            context.Return(result);
+
+            // Compile and return the function.
+
+            ControlFlowGraph cfg = context.GetControlFlowGraph();
+
+            OperandType[] argTypes = new OperandType[] { OperandType.I64 };
+
+            return Compiler.Compile(cfg, argTypes, OperandType.I32, CompilerOptions.HighCq).Map<DebugThreadLocalMapGetOrReserve>();
+        }
+
+        public static DebugNativeWriteLoop GenerateDebugNativeWriteLoop()
+        {
+            EmitterContext context = new EmitterContext();
+
+            // Loop a write to the target address until "running" is false.
+
+            Operand structPtr = context.Copy(context.LoadArgument(OperandType.I64, 0));
+            Operand writePtr = context.Copy(context.LoadArgument(OperandType.I64, 1));
+
+            Operand loopLabel = Label();
+            context.MarkLabel(loopLabel);
+
+            context.Store(writePtr, Const(12345));
+
+            Operand running = context.Load(OperandType.I32, structPtr);
+
+            context.BranchIfTrue(loopLabel, running);
+
+            context.Return();
+
+            // Compile and return the function.
+
+            ControlFlowGraph cfg = context.GetControlFlowGraph();
+
+            OperandType[] argTypes = new OperandType[] { OperandType.I64 };
+
+            return Compiler.Compile(cfg, argTypes, OperandType.None, CompilerOptions.HighCq).Map<DebugNativeWriteLoop>();
+        }
+    }
+}

ARMeilleure/Signal/WindowsPartialUnmapHandler.cs

@@ -0,0 +1,186 @@
+using ARMeilleure.IntermediateRepresentation;
+using ARMeilleure.Translation;
+using Ryujinx.Common.Memory.PartialUnmaps;
+using System;
+
+using static ARMeilleure.IntermediateRepresentation.Operand.Factory;
+
+namespace ARMeilleure.Signal
+{
+    /// <summary>
+    /// Methods to handle signals caused by partial unmaps. See the structs for C# implementations of the methods.
+    /// </summary>
+    internal static class WindowsPartialUnmapHandler
+    {
+        public static Operand EmitRetryFromAccessViolation(EmitterContext context)
+        {
+            IntPtr partialRemapStatePtr = PartialUnmapState.GlobalState;
+            IntPtr localCountsPtr = IntPtr.Add(partialRemapStatePtr, PartialUnmapState.LocalCountsOffset);
+
+            // Get the lock first.
+            EmitNativeReaderLockAcquire(context, IntPtr.Add(partialRemapStatePtr, PartialUnmapState.PartialUnmapLockOffset));
+
+            IntPtr getCurrentThreadId = WindowsSignalHandlerRegistration.GetCurrentThreadIdFunc();
+            Operand threadId = context.Call(Const((ulong)getCurrentThreadId), OperandType.I32);
+            Operand threadIndex = EmitThreadLocalMapIntGetOrReserve(context, localCountsPtr, threadId, Const(0));
+
+            Operand endLabel = Label();
+            Operand retry = context.AllocateLocal(OperandType.I32);
+            Operand threadIndexValidLabel = Label();
+
+            context.BranchIfFalse(threadIndexValidLabel, context.ICompareEqual(threadIndex, Const(-1)));
+
+            context.Copy(retry, Const(1)); // Always retry when thread local cannot be allocated.
+
+            context.Branch(endLabel);
+
+            context.MarkLabel(threadIndexValidLabel);
+
+            Operand threadLocalPartialUnmapsPtr = EmitThreadLocalMapIntGetValuePtr(context, localCountsPtr, threadIndex);
+            Operand threadLocalPartialUnmaps = context.Load(OperandType.I32, threadLocalPartialUnmapsPtr);
+            Operand partialUnmapsCount = context.Load(OperandType.I32, Const((ulong)IntPtr.Add(partialRemapStatePtr, PartialUnmapState.PartialUnmapsCountOffset)));
+
+            context.Copy(retry, context.ICompareNotEqual(threadLocalPartialUnmaps, partialUnmapsCount));
+
+            Operand noRetryLabel = Label();
+
+            context.BranchIfFalse(noRetryLabel, retry);
+
+            // if (retry) {
+
+            context.Store(threadLocalPartialUnmapsPtr, partialUnmapsCount);
+
+            context.Branch(endLabel);
+
+            context.MarkLabel(noRetryLabel);
+
+            // }
+
+            context.MarkLabel(endLabel);
+
+            // Finally, release the lock and return the retry value.
+            EmitNativeReaderLockRelease(context, IntPtr.Add(partialRemapStatePtr, PartialUnmapState.PartialUnmapLockOffset));
+
+            return retry;
+        }
+
+        public static Operand EmitThreadLocalMapIntGetOrReserve(EmitterContext context, IntPtr threadLocalMapPtr, Operand threadId, Operand initialState)
+        {
+            Operand idsPtr = Const((ulong)IntPtr.Add(threadLocalMapPtr, ThreadLocalMap<int>.ThreadIdsOffset));
+
+            Operand i = context.AllocateLocal(OperandType.I32);
+
+            context.Copy(i, Const(0));
+
+            // (Loop 1) Check all slots for a matching Thread ID (while also trying to allocate)
+
+            Operand endLabel = Label();
+
+            Operand loopLabel = Label();
+            context.MarkLabel(loopLabel);
+
+            Operand offset = context.Multiply(i, Const(sizeof(int)));
+            Operand idPtr = context.Add(idsPtr, context.SignExtend32(OperandType.I64, offset));
+
+            // Check that this slot has the thread ID.
+            Operand existingId = context.CompareAndSwap(idPtr, threadId, threadId);
+
+            // If it was already the thread ID, then we just need to return i.
+            context.BranchIfTrue(endLabel, context.ICompareEqual(existingId, threadId));
+
+            context.Copy(i, context.Add(i, Const(1)));
+
+            context.BranchIfTrue(loopLabel, context.ICompareLess(i, Const(ThreadLocalMap<int>.MapSize)));
+
+            // (Loop 2) Try take a slot that is 0 with our Thread ID.
+
+            context.Copy(i, Const(0)); // Reset i.
+
+            Operand loop2Label = Label();
+            context.MarkLabel(loop2Label);
+
+            Operand offset2 = context.Multiply(i, Const(sizeof(int)));
+            Operand idPtr2 = context.Add(idsPtr, context.SignExtend32(OperandType.I64, offset2));
+
+            // Try and swap in the thread id on top of 0.
+            Operand existingId2 = context.CompareAndSwap(idPtr2, Const(0), threadId);
+
+            Operand idNot0Label = Label();
+
+            // If it was 0, then we need to initialize the struct entry and return i.
+            context.BranchIfFalse(idNot0Label, context.ICompareEqual(existingId2, Const(0)));
+
+            Operand structsPtr = Const((ulong)IntPtr.Add(threadLocalMapPtr, ThreadLocalMap<int>.StructsOffset));
+            Operand structPtr = context.Add(structsPtr, context.SignExtend32(OperandType.I64, offset2));
+            context.Store(structPtr, initialState);
+
+            context.Branch(endLabel);
+
+            context.MarkLabel(idNot0Label);
+
+            context.Copy(i, context.Add(i, Const(1)));
+
+            context.BranchIfTrue(loop2Label, context.ICompareLess(i, Const(ThreadLocalMap<int>.MapSize)));
+
+            context.Copy(i, Const(-1)); // Could not place the thread in the list.
+
+            context.MarkLabel(endLabel);
+
+            return context.Copy(i);
+        }
+
+        private static Operand EmitThreadLocalMapIntGetValuePtr(EmitterContext context, IntPtr threadLocalMapPtr, Operand index)
+        {
+            Operand offset = context.Multiply(index, Const(sizeof(int)));
+            Operand structsPtr = Const((ulong)IntPtr.Add(threadLocalMapPtr, ThreadLocalMap<int>.StructsOffset));
+
+            return context.Add(structsPtr, context.SignExtend32(OperandType.I64, offset));
+        }
+
+        private static void EmitThreadLocalMapIntRelease(EmitterContext context, IntPtr threadLocalMapPtr, Operand threadId, Operand index)
+        {
+            Operand offset = context.Multiply(index, Const(sizeof(int)));
+            Operand idsPtr = Const((ulong)IntPtr.Add(threadLocalMapPtr, ThreadLocalMap<int>.ThreadIdsOffset));
+            Operand idPtr = context.Add(idsPtr, context.SignExtend32(OperandType.I64, offset));
+
+            context.CompareAndSwap(idPtr, threadId, Const(0));
+        }
+
+        private static void EmitAtomicAddI32(EmitterContext context, Operand ptr, Operand additive)
+        {
+            Operand loop = Label();
+            context.MarkLabel(loop);
+
+            Operand initial = context.Load(OperandType.I32, ptr);
+            Operand newValue = context.Add(initial, additive);
+
+            Operand replaced = context.CompareAndSwap(ptr, initial, newValue);
+
+            context.BranchIfFalse(loop, context.ICompareEqual(initial, replaced));
+        }
+
+        private static void EmitNativeReaderLockAcquire(EmitterContext context, IntPtr nativeReaderLockPtr)
+        {
+            Operand writeLockPtr = Const((ulong)IntPtr.Add(nativeReaderLockPtr, NativeReaderWriterLock.WriteLockOffset));
+
+            // Spin until we can acquire the write lock.
+            Operand spinLabel = Label();
+            context.MarkLabel(spinLabel);
+
+            // Old value must be 0 to continue (we gained the write lock)
+            context.BranchIfTrue(spinLabel, context.CompareAndSwap(writeLockPtr, Const(0), Const(1)));
+
+            // Increment reader count.
+            EmitAtomicAddI32(context, Const((ulong)IntPtr.Add(nativeReaderLockPtr, NativeReaderWriterLock.ReaderCountOffset)), Const(1));
+
+            // Release write lock.
+            context.CompareAndSwap(writeLockPtr, Const(1), Const(0));
+        }
+
+        private static void EmitNativeReaderLockRelease(EmitterContext context, IntPtr nativeReaderLockPtr)
+        {
+            // Decrement reader count.
+            EmitAtomicAddI32(context, Const((ulong)IntPtr.Add(nativeReaderLockPtr, NativeReaderWriterLock.ReaderCountOffset)), Const(-1));
+        }
+    }
+}

ARMeilleure/Signal/WindowsSignalHandlerRegistration.cs

@@ -3,7 +3,7 @@ using System.Runtime.InteropServices;
 
 namespace ARMeilleure.Signal
 {
-    class WindowsSignalHandlerRegistration
+    unsafe class WindowsSignalHandlerRegistration
     {
         [DllImport("kernel32.dll")]
         private static extern IntPtr AddVectoredExceptionHandler(uint first, IntPtr handler);
@@ -11,6 +11,14 @@ namespace ARMeilleure.Signal
         [DllImport("kernel32.dll")]
         private static extern ulong RemoveVectoredExceptionHandler(IntPtr handle);
 
+        [DllImport("kernel32.dll", SetLastError = true, CharSet = CharSet.Ansi)]
+        static extern IntPtr LoadLibrary([MarshalAs(UnmanagedType.LPStr)] string lpFileName);
+
+        [DllImport("kernel32.dll", CharSet = CharSet.Ansi, ExactSpelling = true, SetLastError = true)]
+        private static extern IntPtr GetProcAddress(IntPtr hModule, string procName);
+
+        private static IntPtr _getCurrentThreadIdPtr;
+
         public static IntPtr RegisterExceptionHandler(IntPtr action)
         {
             return AddVectoredExceptionHandler(1, action);
@@ -20,5 +28,17 @@ namespace ARMeilleure.Signal
         {
             return RemoveVectoredExceptionHandler(handle) != 0;
         }
+
+        public static IntPtr GetCurrentThreadIdFunc()
+        {
+            if (_getCurrentThreadIdPtr == IntPtr.Zero)
+            {
+                IntPtr handle = LoadLibrary("kernel32.dll");
+
+                _getCurrentThreadIdPtr = GetProcAddress(handle, "GetCurrentThreadId");
+            }
+
+            return _getCurrentThreadIdPtr;
+        }
     }
 }

ARMeilleure/Translation/Delegates.cs

@@ -127,7 +127,7 @@ namespace ARMeilleure.Translation
             SetDelegateInfo(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpcr)));
             SetDelegateInfo(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpscr))); // A32 only.
             SetDelegateInfo(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsr)));
-            SetDelegateInfo(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc))); // A32 only.
+            SetDelegateInfo(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetFpsrQc)));
             SetDelegateInfo(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetTpidrEl0)));
             SetDelegateInfo(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SetTpidrEl032))); // A32 only.
             SetDelegateInfo(typeof(NativeInterface).GetMethod(nameof(NativeInterface.SignalMemoryTracking)));
@@ -140,12 +140,6 @@ namespace ARMeilleure.Translation
             SetDelegateInfo(typeof(NativeInterface).GetMethod(nameof(NativeInterface.WriteUInt64)));
             SetDelegateInfo(typeof(NativeInterface).GetMethod(nameof(NativeInterface.WriteVector128)));
 
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinarySignedSatQAcc)));
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinarySignedSatQAdd)));
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinarySignedSatQSub)));
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinaryUnsignedSatQAcc)));
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinaryUnsignedSatQAdd)));
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.BinaryUnsignedSatQSub)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.CountLeadingSigns)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.CountLeadingZeros)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Crc32b)));
@@ -188,8 +182,6 @@ namespace ARMeilleure.Translation
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.SignedShlReg)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.SignedShlRegSatQ)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.SignedShrImm64)));
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.SignedSrcSignedDstSatQ)));
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.SignedSrcUnsignedDstSatQ)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Tbl1)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Tbl2)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Tbl3)));
@@ -198,12 +190,9 @@ namespace ARMeilleure.Translation
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Tbx2)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Tbx3)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.Tbx4)));
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.UnarySignedSatQAbsOrNeg)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.UnsignedShlReg)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.UnsignedShlRegSatQ)));
             SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.UnsignedShrImm64)));
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.UnsignedSrcSignedDstSatQ)));
-            SetDelegateInfo(typeof(SoftFallback).GetMethod(nameof(SoftFallback.UnsignedSrcUnsignedDstSatQ)));
 
             SetDelegateInfo(typeof(SoftFloat16_32).GetMethod(nameof(SoftFloat16_32.FPConvert)));
             SetDelegateInfo(typeof(SoftFloat16_64).GetMethod(nameof(SoftFloat16_64.FPConvert)));

ARMeilleure/Translation/IntervalTree.cs

@@ -19,8 +19,6 @@ namespace ARMeilleure.Translation
 
         public int Count => _count;
 
-        public IntervalTree() { }
-
         #region Public Methods
 
         /// <summary>
@@ -344,7 +342,7 @@ namespace ARMeilleure.Translation
         }
 
         /// <summary>
-        /// Removes the value from the dictionary after searching for it with <paramref name="key">.
+        /// Removes the value from the dictionary after searching for it with <paramref name="key"/>.
         /// </summary>
         /// <param name="key">Key to search for</param>
         /// <returns>Number of deleted values</returns>

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 = 3439; //! To be incremented manually for each change to the ARMeilleure project.
+        private const uint InternalVersion = 3683; //! To be incremented manually for each change to the ARMeilleure project.
 
         private const string ActualDir = "0";
         private const string BackupDir = "1";
@@ -946,9 +946,12 @@ namespace ARMeilleure.Translation.PTC
             return BitConverter.IsLittleEndian;
         }
 
-        private static ulong GetFeatureInfo()
+        private static FeatureInfo GetFeatureInfo()
         {
-            return (ulong)HardwareCapabilities.FeatureInfoEdx << 32 | (uint)HardwareCapabilities.FeatureInfoEcx;
+            return new FeatureInfo(
+                (uint)HardwareCapabilities.FeatureInfo1Ecx,
+                (uint)HardwareCapabilities.FeatureInfo1Edx,
+                (uint)HardwareCapabilities.FeatureInfo7Ebx);
         }
 
         private static byte GetMemoryManagerMode()
@@ -968,7 +971,7 @@ namespace ARMeilleure.Translation.PTC
             return osPlatform;
         }
 
-        [StructLayout(LayoutKind.Sequential, Pack = 1/*, Size = 50*/)]
+        [StructLayout(LayoutKind.Sequential, Pack = 1/*, Size = 54*/)]
         private struct OuterHeader
         {
             public ulong Magic;
@@ -976,7 +979,7 @@ namespace ARMeilleure.Translation.PTC
             public uint CacheFileVersion;
 
             public bool Endianness;
-            public ulong FeatureInfo;
+            public FeatureInfo FeatureInfo;
             public byte MemoryManagerMode;
             public uint OSPlatform;
 
@@ -999,6 +1002,9 @@ namespace ARMeilleure.Translation.PTC
             }
         }
 
+        [StructLayout(LayoutKind.Sequential, Pack = 1/*, Size = 12*/)]
+        private record struct FeatureInfo(uint FeatureInfo0, uint FeatureInfo1, uint FeatureInfo2);
+
         [StructLayout(LayoutKind.Sequential, Pack = 1/*, Size = 128*/)]
         private struct InnerHeader
         {

ARMeilleure/Translation/Translator.cs

@@ -44,15 +44,11 @@ namespace ARMeilleure.Translation
         private readonly IJitMemoryAllocator _allocator;
         private readonly ConcurrentQueue<KeyValuePair<ulong, TranslatedFunction>> _oldFuncs;
 
-        private readonly ConcurrentDictionary<ulong, object> _backgroundSet;
-        private readonly ConcurrentStack<RejitRequest> _backgroundStack;
-        private readonly AutoResetEvent _backgroundTranslatorEvent;
-        private readonly ReaderWriterLock _backgroundTranslatorLock;
-
         internal TranslatorCache<TranslatedFunction> Functions { get; }
         internal AddressTable<ulong> FunctionTable { get; }
         internal EntryTable<uint> CountTable { get; }
         internal TranslatorStubs Stubs { get; }
+        internal TranslatorQueue Queue { get; }
         internal IMemoryManager Memory { get; }
 
         private volatile int _threadCount;
@@ -67,10 +63,7 @@ namespace ARMeilleure.Translation
 
             _oldFuncs = new ConcurrentQueue<KeyValuePair<ulong, TranslatedFunction>>();
 
-            _backgroundSet = new ConcurrentDictionary<ulong, object>();
-            _backgroundStack = new ConcurrentStack<RejitRequest>();
-            _backgroundTranslatorEvent = new AutoResetEvent(false);
-            _backgroundTranslatorLock = new ReaderWriterLock();
+            Queue = new TranslatorQueue();
 
             JitCache.Initialize(allocator);
 
@@ -87,43 +80,6 @@ namespace ARMeilleure.Translation
             }
         }
 
-        private void TranslateStackedSubs()
-        {
-            while (_threadCount != 0)
-            {
-                _backgroundTranslatorLock.AcquireReaderLock(Timeout.Infinite);
-
-                if (_backgroundStack.TryPop(out RejitRequest request) &&
-                    _backgroundSet.TryRemove(request.Address, out _))
-                {
-                    TranslatedFunction func = Translate(request.Address, request.Mode, highCq: true);
-
-                    Functions.AddOrUpdate(request.Address, func.GuestSize, func, (key, oldFunc) =>
-                    {
-                        EnqueueForDeletion(key, oldFunc);
-                        return func;
-                    });
-
-                    if (PtcProfiler.Enabled)
-                    {
-                        PtcProfiler.UpdateEntry(request.Address, request.Mode, highCq: true);
-                    }
-
-                    RegisterFunction(request.Address, func);
-
-                    _backgroundTranslatorLock.ReleaseReaderLock();
-                }
-                else
-                {
-                    _backgroundTranslatorLock.ReleaseReaderLock();
-                    _backgroundTranslatorEvent.WaitOne();
-                }
-            }
-
-             // Wake up any other background translator threads, to encourage them to exit.
-            _backgroundTranslatorEvent.Set();
-        }
-
         public void Execute(State.ExecutionContext context, ulong address)
         {
             if (Interlocked.Increment(ref _threadCount) == 1)
@@ -155,7 +111,7 @@ namespace ARMeilleure.Translation
                 {
                     bool last = i != 0 && i == unboundedThreadCount - 1;
 
-                    Thread backgroundTranslatorThread = new Thread(TranslateStackedSubs)
+                    Thread backgroundTranslatorThread = new Thread(BackgroundTranslate)
                     {
                         Name = "CPU.BackgroundTranslatorThread." + i,
                         Priority = last ? ThreadPriority.Lowest : ThreadPriority.Normal
@@ -186,10 +142,9 @@ namespace ARMeilleure.Translation
 
             if (Interlocked.Decrement(ref _threadCount) == 0)
             {
-                _backgroundTranslatorEvent.Set();
-
                 ClearJitCache();
 
+                Queue.Dispose();
                 Stubs.Dispose();
                 FunctionTable.Dispose();
                 CountTable.Dispose();
@@ -317,6 +272,27 @@ namespace ARMeilleure.Translation
             return new TranslatedFunction(func, counter, funcSize, highCq);
         }
 
+        private void BackgroundTranslate()
+        {
+            while (_threadCount != 0 && Queue.TryDequeue(out RejitRequest request))
+            {
+                TranslatedFunction func = Translate(request.Address, request.Mode, highCq: true);
+
+                Functions.AddOrUpdate(request.Address, func.GuestSize, func, (key, oldFunc) =>
+                {
+                    EnqueueForDeletion(key, oldFunc);
+                    return func;
+                });
+
+                if (PtcProfiler.Enabled)
+                {
+                    PtcProfiler.UpdateEntry(request.Address, request.Mode, highCq: true);
+                }
+
+                RegisterFunction(request.Address, func);
+            }
+        }
+
         private struct Range
         {
             public ulong Start { get; }
@@ -504,11 +480,7 @@ namespace ARMeilleure.Translation
 
         internal void EnqueueForRejit(ulong guestAddress, ExecutionMode mode)
         {
-            if (_backgroundSet.TryAdd(guestAddress, null))
-            {
-                _backgroundStack.Push(new RejitRequest(guestAddress, mode));
-                _backgroundTranslatorEvent.Set();
-            }
+            Queue.Enqueue(guestAddress, mode);
         }
 
         private void EnqueueForDeletion(ulong guestAddress, TranslatedFunction func)
@@ -542,26 +514,23 @@ namespace ARMeilleure.Translation
 
         private void ClearRejitQueue(bool allowRequeue)
         {
-            _backgroundTranslatorLock.AcquireWriterLock(Timeout.Infinite);
-
-            if (allowRequeue)
+            if (!allowRequeue)
             {
-                while (_backgroundStack.TryPop(out var request))
+                Queue.Clear();
+
+                return;
+            }
+
+            lock (Queue.Sync)
+            {
+                while (Queue.Count > 0 && Queue.TryDequeue(out RejitRequest request))
                 {
                     if (Functions.TryGetValue(request.Address, out var func) && func.CallCounter != null)
                     {
                         Volatile.Write(ref func.CallCounter.Value, 0);
                     }
-
-                    _backgroundSet.TryRemove(request.Address, out _);
                 }
             }
-            else
-            {
-                _backgroundStack.Clear();
-            }
-
-            _backgroundTranslatorLock.ReleaseWriterLock();
         }
     }
 }

ARMeilleure/Translation/TranslatorQueue.cs

@@ -0,0 +1,121 @@
+using ARMeilleure.Diagnostics;
+using ARMeilleure.State;
+using System;
+using System.Collections.Generic;
+using System.Threading;
+
+namespace ARMeilleure.Translation
+{
+    /// <summary>
+    /// Represents a queue of <see cref="RejitRequest"/>.
+    /// </summary>
+    /// <remarks>
+    /// This does not necessarily behave like a queue, i.e: a FIFO collection.
+    /// </remarks>
+    sealed class TranslatorQueue : IDisposable
+    {
+        private bool _disposed;
+        private readonly Stack<RejitRequest> _requests;
+        private readonly HashSet<ulong> _requestAddresses;
+
+        /// <summary>
+        /// Gets the object used to synchronize access to the <see cref="TranslatorQueue"/>.
+        /// </summary>
+        public object Sync { get; }
+
+        /// <summary>
+        /// Gets the number of requests in the <see cref="TranslatorQueue"/>.
+        /// </summary>
+        public int Count => _requests.Count;
+
+        /// <summary>
+        /// Initializes a new instance of the <see cref="TranslatorQueue"/> class.
+        /// </summary>
+        public TranslatorQueue()
+        {
+            Sync = new object();
+
+            _requests = new Stack<RejitRequest>();
+            _requestAddresses = new HashSet<ulong>();
+        }
+
+        /// <summary>
+        /// Enqueues a request with the specified <paramref name="address"/> and <paramref name="mode"/>.
+        /// </summary>
+        /// <param name="address">Address of request</param>
+        /// <param name="mode"><see cref="ExecutionMode"/> of request</param>
+        public void Enqueue(ulong address, ExecutionMode mode)
+        {
+            lock (Sync)
+            {
+                if (_requestAddresses.Add(address))
+                {
+                    _requests.Push(new RejitRequest(address, mode));
+
+                    TranslatorEventSource.Log.RejitQueueAdd(1);
+
+                    Monitor.Pulse(Sync);
+                }
+            }
+        }
+
+        /// <summary>
+        /// Tries to dequeue a <see cref="RejitRequest"/>. This will block the thread until a <see cref="RejitRequest"/>
+        /// is enqueued or the <see cref="TranslatorQueue"/> is disposed.
+        /// </summary>
+        /// <param name="result"><see cref="RejitRequest"/> dequeued</param>
+        /// <returns><see langword="true"/> on success; otherwise <see langword="false"/></returns>
+        public bool TryDequeue(out RejitRequest result)
+        {
+            while (!_disposed)
+            {
+                lock (Sync)
+                {
+                    if (_requests.TryPop(out result))
+                    {
+                        _requestAddresses.Remove(result.Address);
+
+                        TranslatorEventSource.Log.RejitQueueAdd(-1);
+
+                        return true;
+                    }
+
+                    Monitor.Wait(Sync);
+                }
+            }
+
+            result = default;
+
+            return false;
+        }
+
+        /// <summary>
+        /// Clears the <see cref="TranslatorQueue"/>.
+        /// </summary>
+        public void Clear()
+        {
+            lock (Sync)
+            {
+                TranslatorEventSource.Log.RejitQueueAdd(-_requests.Count);
+
+                _requests.Clear();
+                _requestAddresses.Clear();
+
+                Monitor.PulseAll(Sync);
+            }
+        }
+
+        /// <summary>
+        /// Releases all resources used by the <see cref="TranslatorQueue"/> instance.
+        /// </summary>
+        public void Dispose()
+        {
+            if (!_disposed)
+            {
+                _disposed = true;
+
+                Clear();
+            }
+        }
+    }
+}

Ryujinx.Audio.Backends.SDL2/SDL2HardwareDeviceDriver.cs

@@ -4,7 +4,6 @@ using Ryujinx.Memory;
 using Ryujinx.SDL2.Common;
 using System;
 using System.Collections.Concurrent;
-using System.Runtime.InteropServices;
 using System.Threading;
 
 using static Ryujinx.Audio.Integration.IHardwareDeviceDriver;

Ryujinx.Audio.Backends.SDL2/SDL2HardwareDeviceSession.cs

@@ -4,7 +4,6 @@ using Ryujinx.Common.Logging;
 using Ryujinx.Memory;
 using System;
 using System.Collections.Concurrent;
-using System.Runtime.InteropServices;
 using System.Threading;
 
 using static SDL2.SDL;

Ryujinx.Audio.Backends.SoundIo/Native/libsoundio/SoundIOSampleRateRange.cs

@@ -1,5 +1,4 @@
-using System;
-namespace SoundIOSharp
+namespace SoundIOSharp
 {
     public struct SoundIOSampleRateRange
     {

Ryujinx.Audio/Renderer/Common/WorkBufferAllocator.cs

@@ -41,7 +41,7 @@ namespace Ryujinx.Audio.Renderer.Common
             return Memory<byte>.Empty;
         }
 
-        public Memory<T> Allocate<T>(ulong count, int align) where T: unmanaged
+        public Memory<T> Allocate<T>(ulong count, int align) where T : unmanaged
         {
             Memory<byte> allocatedMemory = Allocate((ulong)Unsafe.SizeOf<T>() * count, align);
 
@@ -53,7 +53,7 @@ namespace Ryujinx.Audio.Renderer.Common
             return SpanMemoryManager<T>.Cast(allocatedMemory);
         }
 
-        public static ulong GetTargetSize<T>(ulong currentSize, ulong count, int align) where T: unmanaged
+        public static ulong GetTargetSize<T>(ulong currentSize, ulong count, int align) where T : unmanaged
         {
             return BitUtils.AlignUp(currentSize, align) + (ulong)Unsafe.SizeOf<T>() * count;
         }

Ryujinx.Audio/Renderer/Dsp/Command/BiquadFilterCommand.cs

@@ -1,7 +1,6 @@
 using Ryujinx.Audio.Renderer.Dsp.State;
 using Ryujinx.Audio.Renderer.Parameter;
 using System;
-using System.Runtime.CompilerServices;
 
 namespace Ryujinx.Audio.Renderer.Dsp.Command
 {

Ryujinx.Audio/Renderer/Dsp/Command/CaptureBufferCommand.cs

@@ -1,5 +1,4 @@
 using Ryujinx.Audio.Renderer.Dsp.State;
-using Ryujinx.Audio.Renderer.Parameter;
 using Ryujinx.Memory;
 using System;
 using System.Runtime.CompilerServices;

Ryujinx.Audio/Renderer/Dsp/Command/CircularBufferSinkCommand.cs

@@ -1,6 +1,5 @@
 using Ryujinx.Audio.Renderer.Parameter.Sink;
 using Ryujinx.Audio.Renderer.Server.MemoryPool;
-using System;
 using System.Diagnostics;
 
 namespace Ryujinx.Audio.Renderer.Dsp.Command

Ryujinx.Audio/Renderer/Dsp/Command/CommandList.cs

@@ -6,7 +6,6 @@ using Ryujinx.Memory;
 using System;
 using System.Buffers;
 using System.Collections.Generic;
-using System.Diagnostics;
 using System.Runtime.CompilerServices;
 
 namespace Ryujinx.Audio.Renderer.Dsp.Command

Ryujinx.Audio/Renderer/Dsp/Command/CopyMixBufferCommand.cs

@@ -1,5 +1,3 @@
-using System;
-
 namespace Ryujinx.Audio.Renderer.Dsp.Command
 {
     public class CopyMixBufferCommand : ICommand

Ryujinx.Audio/Renderer/Dsp/Command/DelayCommand.cs

@@ -87,7 +87,7 @@ namespace Ryujinx.Audio.Renderer.Dsp.Command
             float dryGain = FixedPointHelper.ToFloat(Parameter.DryGain, FixedPointPrecision);
             float outGain = FixedPointHelper.ToFloat(Parameter.OutGain, FixedPointPrecision);
 
-            Matrix2x2 delayFeedback = new Matrix2x2(delayFeedbackBaseGain , delayFeedbackCrossGain,
+            Matrix2x2 delayFeedback = new Matrix2x2(delayFeedbackBaseGain, delayFeedbackCrossGain,
                                                     delayFeedbackCrossGain, delayFeedbackBaseGain);
 
             for (int i = 0; i < sampleCount; i++)
@@ -124,10 +124,10 @@ namespace Ryujinx.Audio.Renderer.Dsp.Command
             float dryGain = FixedPointHelper.ToFloat(Parameter.DryGain, FixedPointPrecision);
             float outGain = FixedPointHelper.ToFloat(Parameter.OutGain, FixedPointPrecision);
 
-            Matrix4x4 delayFeedback = new Matrix4x4(delayFeedbackBaseGain , delayFeedbackCrossGain, delayFeedbackCrossGain, 0.0f,
-                                                    delayFeedbackCrossGain, delayFeedbackBaseGain , 0.0f                  , delayFeedbackCrossGain,
-                                                    delayFeedbackCrossGain, 0.0f                  , delayFeedbackBaseGain , delayFeedbackCrossGain,
-                                                    0.0f                  , delayFeedbackCrossGain, delayFeedbackCrossGain, delayFeedbackBaseGain);
+            Matrix4x4 delayFeedback = new Matrix4x4(delayFeedbackBaseGain, delayFeedbackCrossGain, delayFeedbackCrossGain, 0.0f,
+                                                    delayFeedbackCrossGain, delayFeedbackBaseGain, 0.0f, delayFeedbackCrossGain,
+                                                    delayFeedbackCrossGain, 0.0f, delayFeedbackBaseGain, delayFeedbackCrossGain,
+                                                    0.0f, delayFeedbackCrossGain, delayFeedbackCrossGain, delayFeedbackBaseGain);
 
 
             for (int i = 0; i < sampleCount; i++)
@@ -171,12 +171,12 @@ namespace Ryujinx.Audio.Renderer.Dsp.Command
             float dryGain = FixedPointHelper.ToFloat(Parameter.DryGain, FixedPointPrecision);
             float outGain = FixedPointHelper.ToFloat(Parameter.OutGain, FixedPointPrecision);
 
-            Matrix6x6 delayFeedback = new Matrix6x6(delayFeedbackBaseGain , 0.0f                  , 0.0f                  , 0.0f                  , delayFeedbackCrossGain, delayFeedbackCrossGain,
-                                                    0.0f                  , delayFeedbackBaseGain , 0.0f                  , delayFeedbackCrossGain, delayFeedbackCrossGain, 0.0f                  ,
-                                                    delayFeedbackCrossGain, 0.0f                  , delayFeedbackBaseGain , delayFeedbackCrossGain, 0.0f                  , 0.0f                  ,
-                                                    0.0f                  , delayFeedbackCrossGain, delayFeedbackCrossGain, delayFeedbackBaseGain , 0.0f                  , 0.0f                  ,
-                                                    delayFeedbackCrossGain, delayFeedbackCrossGain, 0.0f                  , 0.0f                  , delayFeedbackBaseGain , 0.0f                  ,
-                                                    0.0f                  , 0.0f                  , 0.0f                  , 0.0f                  , 0.0f                  , feedbackGain);
+            Matrix6x6 delayFeedback = new Matrix6x6(delayFeedbackBaseGain, 0.0f, 0.0f, 0.0f, delayFeedbackCrossGain, delayFeedbackCrossGain,
+                                                    0.0f, delayFeedbackBaseGain, 0.0f, delayFeedbackCrossGain, delayFeedbackCrossGain, 0.0f,
+                                                    delayFeedbackCrossGain, 0.0f, delayFeedbackBaseGain, delayFeedbackCrossGain, 0.0f, 0.0f,
+                                                    0.0f, delayFeedbackCrossGain, delayFeedbackCrossGain, delayFeedbackBaseGain, 0.0f, 0.0f,
+                                                    delayFeedbackCrossGain, delayFeedbackCrossGain, 0.0f, 0.0f, delayFeedbackBaseGain, 0.0f,
+                                                    0.0f, 0.0f, 0.0f, 0.0f, 0.0f, feedbackGain);
 
             for (int i = 0; i < sampleCount; i++)
             {

Ryujinx.Audio/Renderer/Dsp/Command/DepopForMixBuffersCommand.cs

@@ -1,5 +1,4 @@
 using System;
-using System.Diagnostics;
 using System.Runtime.CompilerServices;
 
 namespace Ryujinx.Audio.Renderer.Dsp.Command

Ryujinx.Audio/Renderer/Dsp/Command/LimiterCommandVersion1.cs

@@ -2,7 +2,6 @@ using Ryujinx.Audio.Renderer.Dsp.State;
 using Ryujinx.Audio.Renderer.Parameter.Effect;
 using System;
 using System.Diagnostics;
-using System.Runtime.CompilerServices;
 
 namespace Ryujinx.Audio.Renderer.Dsp.Command
 {

Ryujinx.Audio/Renderer/Dsp/Command/LimiterCommandVersion2.cs

@@ -3,7 +3,6 @@ using Ryujinx.Audio.Renderer.Parameter;
 using Ryujinx.Audio.Renderer.Parameter.Effect;
 using System;
 using System.Diagnostics;
-using System.Runtime.CompilerServices;
 using System.Runtime.InteropServices;
 
 namespace Ryujinx.Audio.Renderer.Dsp.Command

Ryujinx.Audio/Renderer/Dsp/Command/UpsampleCommand.cs

@@ -1,6 +1,5 @@
 using Ryujinx.Audio.Renderer.Server.Upsampler;
 using System;
-using System.Runtime.CompilerServices;
 
 namespace Ryujinx.Audio.Renderer.Dsp.Command
 {

Ryujinx.Audio/Renderer/Dsp/DataSourceHelper.cs

@@ -76,7 +76,7 @@ namespace Ryujinx.Audio.Renderer.Dsp
 
                     if (!info.DecodingBehaviour.HasFlag(DecodingBehaviour.SkipPitchAndSampleRateConversion))
                     {
-                        voiceState.Pitch.ToSpan().Slice(0, pitchMaxLength).CopyTo(tempBuffer);
+                        voiceState.Pitch.AsSpan().Slice(0, pitchMaxLength).CopyTo(tempBuffer);
                         tempBufferIndex += pitchMaxLength;
                     }
 
@@ -239,7 +239,7 @@ namespace Ryujinx.Audio.Renderer.Dsp
 
                         ResamplerHelper.Resample(outputBuffer, tempBuffer, sampleRateRatio, ref fraction, sampleCountToProcess, info.SrcQuality, y != sourceSampleCountToProcess || info.Pitch != 1.0f);
 
-                        tempBuffer.Slice(sampleCountToDecode, pitchMaxLength).CopyTo(voiceState.Pitch.ToSpan());
+                        tempBuffer.Slice(sampleCountToDecode, pitchMaxLength).CopyTo(voiceState.Pitch.AsSpan());
                     }
 
                     i += sampleCountToProcess;

Ryujinx.Audio/Renderer/Dsp/State/LimiterState.cs

@@ -25,6 +25,6 @@ namespace Ryujinx.Audio.Renderer.Dsp.State
             UpdateParameter(ref parameter);
         }
 
-        public void UpdateParameter(ref LimiterParameter parameter) {}
+        public void UpdateParameter(ref LimiterParameter parameter) { }
     }
 }

Ryujinx.Audio/Renderer/Parameter/Effect/LimiterStatistics.cs

@@ -24,8 +24,8 @@ namespace Ryujinx.Audio.Renderer.Parameter.Effect
         /// </summary>
         public void Reset()
         {
-            InputMax.ToSpan().Fill(0.0f);
-            CompressionGainMin.ToSpan().Fill(1.0f);
+            InputMax.AsSpan().Fill(0.0f);
+            CompressionGainMin.AsSpan().Fill(1.0f);
         }
     }
 }

Ryujinx.Audio/Renderer/Server/AudioRenderSystem.cs

@@ -55,7 +55,6 @@ namespace Ryujinx.Audio.Renderer.Server
         private uint _processHandle;
         private ulong _appletResourceId;
 
-        private WritableRegion _workBufferRegion;
         private MemoryHandle _workBufferMemoryPin;
 
         private Memory<float> _mixBuffer;
@@ -98,7 +97,15 @@ namespace Ryujinx.Audio.Renderer.Server
             _sessionId = 0;
         }
 
-        public ResultCode Initialize(ref AudioRendererConfiguration parameter, uint processHandle, CpuAddress workBuffer, ulong workBufferSize, int sessionId, ulong appletResourceId, IVirtualMemoryManager memoryManager)
+        public ResultCode Initialize(
+            ref AudioRendererConfiguration parameter,
+            uint processHandle,
+            Memory<byte> workBufferMemory,
+            CpuAddress workBuffer,
+            ulong workBufferSize,
+            int sessionId,
+            ulong appletResourceId,
+            IVirtualMemoryManager memoryManager)
         {
             if (!BehaviourContext.CheckValidRevision(parameter.Revision))
             {
@@ -134,11 +141,10 @@ namespace Ryujinx.Audio.Renderer.Server
 
             WorkBufferAllocator workBufferAllocator;
 
-            _workBufferRegion = MemoryManager.GetWritableRegion(workBuffer, (int)workBufferSize);
-            _workBufferRegion.Memory.Span.Fill(0);
-            _workBufferMemoryPin = _workBufferRegion.Memory.Pin();
+            workBufferMemory.Span.Fill(0);
+            _workBufferMemoryPin = workBufferMemory.Pin();
 
-            workBufferAllocator = new WorkBufferAllocator(_workBufferRegion.Memory);
+            workBufferAllocator = new WorkBufferAllocator(workBufferMemory);
 
             PoolMapper poolMapper = new PoolMapper(processHandle, false);
             poolMapper.InitializeSystemPool(ref _dspMemoryPoolState, workBuffer, workBufferSize);
@@ -841,7 +847,6 @@ namespace Ryujinx.Audio.Renderer.Server
                 _manager.Unregister(this);
                 _terminationEvent.Dispose();
                 _workBufferMemoryPin.Dispose();
-                _workBufferRegion.Dispose();
 
                 if (MemoryManager is IRefCounted rc)
                 {

Ryujinx.Audio/Renderer/Server/AudioRendererManager.cs

@@ -305,13 +305,34 @@ namespace Ryujinx.Audio.Renderer.Server
         /// <param name="workBufferSize">The guest work buffer size.</param>
         /// <param name="processHandle">The process handle of the application.</param>
         /// <returns>A <see cref="ResultCode"/> reporting an error or a success.</returns>
-        public ResultCode OpenAudioRenderer(out AudioRenderSystem renderer, IVirtualMemoryManager memoryManager, ref AudioRendererConfiguration parameter, ulong appletResourceUserId, ulong workBufferAddress, ulong workBufferSize, uint processHandle, float volume)
+        public ResultCode OpenAudioRenderer(
+            out AudioRenderSystem renderer,
+            IVirtualMemoryManager memoryManager,
+            ref AudioRendererConfiguration parameter,
+            ulong appletResourceUserId,
+            ulong workBufferAddress,
+            ulong workBufferSize,
+            uint processHandle,
+            float volume)
         {
             int sessionId = AcquireSessionId();
 
             AudioRenderSystem audioRenderer = new AudioRenderSystem(this, _sessionsSystemEvent[sessionId]);
 
-            ResultCode result = audioRenderer.Initialize(ref parameter, processHandle, workBufferAddress, workBufferSize, sessionId, appletResourceUserId, memoryManager);
+            // TODO: Eventually, we should try to use the guest supplied work buffer instead of allocating
+            // our own. However, it was causing problems on some applications that would unmap the memory
+            // before the audio renderer was fully disposed.
+            Memory<byte> workBufferMemory = GC.AllocateArray<byte>((int)workBufferSize, pinned: true);
+
+            ResultCode result = audioRenderer.Initialize(
+                ref parameter,
+                processHandle,
+                workBufferMemory,
+                workBufferAddress,
+                workBufferSize,
+                sessionId,
+                appletResourceUserId,
+                memoryManager);
 
             if (result == ResultCode.Success)
             {

Ryujinx.Audio/Renderer/Server/CommandBuffer.cs

@@ -7,7 +7,6 @@ using Ryujinx.Audio.Renderer.Server.Performance;
 using Ryujinx.Audio.Renderer.Server.Sink;
 using Ryujinx.Audio.Renderer.Server.Upsampler;
 using Ryujinx.Audio.Renderer.Server.Voice;
-using Ryujinx.Common.Memory;
 using System;
 using CpuAddress = System.UInt64;
 

Ryujinx.Audio/Renderer/Server/CommandGenerator.cs

@@ -141,7 +141,7 @@ namespace Ryujinx.Audio.Renderer.Server
                 Memory<byte> biquadStateRawMemory = SpanMemoryManager<byte>.Cast(state).Slice(VoiceUpdateState.BiquadStateOffset, VoiceUpdateState.BiquadStateSize * Constants.VoiceBiquadFilterCount);
                 Memory<BiquadFilterState> stateMemory = SpanMemoryManager<BiquadFilterState>.Cast(biquadStateRawMemory);
 
-                _commandBuffer.GenerateGroupedBiquadFilter(baseIndex, voiceState.BiquadFilters.ToSpan(), stateMemory, bufferOffset, bufferOffset, voiceState.BiquadFilterNeedInitialization, nodeId);
+                _commandBuffer.GenerateGroupedBiquadFilter(baseIndex, voiceState.BiquadFilters.AsSpan(), stateMemory, bufferOffset, bufferOffset, voiceState.BiquadFilterNeedInitialization, nodeId);
             }
             else
             {
@@ -337,8 +337,8 @@ namespace Ryujinx.Audio.Renderer.Server
                             GeneratePerformance(ref performanceEntry, PerformanceCommand.Type.Start, nodeId);
                         }
 
-                        GenerateVoiceMix(channelResource.Mix.ToSpan(),
-                                         channelResource.PreviousMix.ToSpan(),
+                        GenerateVoiceMix(channelResource.Mix.AsSpan(),
+                                         channelResource.PreviousMix.AsSpan(),
                                          dspStateMemory,
                                          mix.BufferOffset,
                                          mix.BufferCount,
@@ -505,8 +505,8 @@ namespace Ryujinx.Audio.Renderer.Server
                 BiquadFilterParameter parameter = new BiquadFilterParameter();
 
                 parameter.Enable = true;
-                effect.Parameter.Denominator.ToSpan().CopyTo(parameter.Denominator.ToSpan());
-                effect.Parameter.Numerator.ToSpan().CopyTo(parameter.Numerator.ToSpan());
+                effect.Parameter.Denominator.AsSpan().CopyTo(parameter.Denominator.AsSpan());
+                effect.Parameter.Numerator.AsSpan().CopyTo(parameter.Numerator.AsSpan());
 
                 for (int i = 0; i < effect.Parameter.ChannelCount; i++)
                 {
@@ -923,8 +923,8 @@ namespace Ryujinx.Audio.Renderer.Server
             if (useCustomDownMixingCommand)
             {
                 _commandBuffer.GenerateDownMixSurroundToStereo(finalMix.BufferOffset,
-                                                               sink.Parameter.Input.ToSpan(),
-                                                               sink.Parameter.Input.ToSpan(),
+                                                               sink.Parameter.Input.AsSpan(),
+                                                               sink.Parameter.Input.AsSpan(),
                                                                sink.DownMixCoefficients,
                                                                Constants.InvalidNodeId);
             }
@@ -932,8 +932,8 @@ namespace Ryujinx.Audio.Renderer.Server
             else if (_rendererContext.ChannelCount == 2 && sink.Parameter.InputCount == 6)
             {
                 _commandBuffer.GenerateDownMixSurroundToStereo(finalMix.BufferOffset,
-                                                               sink.Parameter.Input.ToSpan(),
-                                                               sink.Parameter.Input.ToSpan(),
+                                                               sink.Parameter.Input.AsSpan(),
+                                                               sink.Parameter.Input.AsSpan(),
                                                                Constants.DefaultSurroundToStereoCoefficients,
                                                                Constants.InvalidNodeId);
             }
@@ -945,7 +945,7 @@ namespace Ryujinx.Audio.Renderer.Server
                 _commandBuffer.GenerateUpsample(finalMix.BufferOffset,
                                                 sink.UpsamplerState,
                                                 sink.Parameter.InputCount,
-                                                sink.Parameter.Input.ToSpan(),
+                                                sink.Parameter.Input.AsSpan(),
                                                 commandList.BufferCount,
                                                 commandList.SampleCount,
                                                 commandList.SampleRate,

Ryujinx.Audio/Renderer/Server/CommandProcessingTimeEstimatorVersion4.cs

@@ -1,9 +1,5 @@
-using Ryujinx.Audio.Common;
 using Ryujinx.Audio.Renderer.Dsp.Command;
-using Ryujinx.Audio.Renderer.Parameter.Effect;
-using System;
 using System.Diagnostics;
-using static Ryujinx.Audio.Renderer.Parameter.VoiceInParameter;
 
 namespace Ryujinx.Audio.Renderer.Server
 {

Ryujinx.Audio/Renderer/Server/Effect/BaseEffect.cs

@@ -81,7 +81,7 @@ namespace Ryujinx.Audio.Renderer.Server.Effect
         /// </summary>
         /// <param name="parameter">The user parameter.</param>
         /// <returns>Returns true if the <see cref="EffectType"/> sent by the user matches the internal <see cref="EffectType"/>.</returns>
-        public bool IsTypeValid<T>(ref T parameter) where T: unmanaged, IEffectInParameter
+        public bool IsTypeValid<T>(ref T parameter) where T : unmanaged, IEffectInParameter
         {
             return parameter.Type == TargetEffectType;
         }
@@ -140,14 +140,14 @@ namespace Ryujinx.Audio.Renderer.Server.Effect
         /// Initialize the given <paramref name="state"/> result state.
         /// </summary>
         /// <param name="state">The state to initalize</param>
-        public virtual void InitializeResultState(ref EffectResultState state) {}
+        public virtual void InitializeResultState(ref EffectResultState state) { }
 
         /// <summary>
         /// Update the <paramref name="destState"/> result state with <paramref name="srcState"/>.
         /// </summary>
         /// <param name="destState">The destination result state</param>
         /// <param name="srcState">The source result state</param>
-        public virtual void UpdateResultState(ref EffectResultState destState, ref EffectResultState srcState) {}
+        public virtual void UpdateResultState(ref EffectResultState destState, ref EffectResultState srcState) { }
 
         /// <summary>
         /// Update the internal state from a user version 1 parameter.
@@ -215,7 +215,7 @@ namespace Ryujinx.Audio.Renderer.Server.Effect
         /// </summary>
         /// <param name="outStatus">The given user output.</param>
         /// <param name="isAudioRendererActive">If set to true, the <see cref="AudioRenderSystem"/> is active.</param>
-        public void StoreStatus<T>(ref T outStatus, bool isAudioRendererActive) where T: unmanaged, IEffectOutStatus
+        public void StoreStatus<T>(ref T outStatus, bool isAudioRendererActive) where T : unmanaged, IEffectOutStatus
         {
             if (isAudioRendererActive)
             {

Ryujinx.Audio/Renderer/Server/Effect/Reverb3dEffect.cs

@@ -46,7 +46,7 @@ namespace Ryujinx.Audio.Renderer.Server.Effect
             Update(out updateErrorInfo, ref parameter, mapper);
         }
 
-        public void Update<T>(out BehaviourParameter.ErrorInfo updateErrorInfo, ref T parameter, PoolMapper mapper) where T: unmanaged, IEffectInParameter
+        public void Update<T>(out BehaviourParameter.ErrorInfo updateErrorInfo, ref T parameter, PoolMapper mapper) where T : unmanaged, IEffectInParameter
         {
             Debug.Assert(IsTypeValid(ref parameter));
 

Ryujinx.Audio/Renderer/Server/MemoryPool/AddressInfo.cs

@@ -1,8 +1,7 @@
 using System;
 using System.Runtime.InteropServices;
-
-using DspAddress = System.UInt64;
 using CpuAddress = System.UInt64;
+using DspAddress = System.UInt64;
 
 namespace Ryujinx.Audio.Renderer.Server.MemoryPool
 {

Ryujinx.Audio/Renderer/Server/MemoryPool/MemoryPoolState.cs

@@ -1,8 +1,7 @@
 using System;
 using System.Runtime.InteropServices;
-
-using DspAddress = System.UInt64;
 using CpuAddress = System.UInt64;
+using DspAddress = System.UInt64;
 
 namespace Ryujinx.Audio.Renderer.Server.MemoryPool
 {

Ryujinx.Audio/Renderer/Server/Performance/PerformanceManagerGeneric.cs

@@ -13,7 +13,7 @@ namespace Ryujinx.Audio.Renderer.Server.Performance
     /// <typeparam name="THeader">The header implementation of the performance frame.</typeparam>
     /// <typeparam name="TEntry">The entry implementation of the performance frame.</typeparam>
     /// <typeparam name="TEntryDetail">A detailed implementation of the performance frame.</typeparam>
-    public class PerformanceManagerGeneric<THeader, TEntry, TEntryDetail> : PerformanceManager where THeader: unmanaged, IPerformanceHeader where TEntry : unmanaged, IPerformanceEntry where TEntryDetail: unmanaged, IPerformanceDetailEntry
+    public class PerformanceManagerGeneric<THeader, TEntry, TEntryDetail> : PerformanceManager where THeader : unmanaged, IPerformanceHeader where TEntry : unmanaged, IPerformanceEntry where TEntryDetail : unmanaged, IPerformanceDetailEntry
     {
         /// <summary>
         /// The magic used for the <see cref="THeader"/>.

Ryujinx.Audio/Renderer/Server/Sink/DeviceSink.cs

@@ -63,10 +63,10 @@ namespace Ryujinx.Audio.Renderer.Server.Sink
             else
             {
                 Parameter.DownMixParameterEnabled = inputDeviceParameter.DownMixParameterEnabled;
-                inputDeviceParameter.DownMixParameter.ToSpan().CopyTo(Parameter.DownMixParameter.ToSpan());
+                inputDeviceParameter.DownMixParameter.AsSpan().CopyTo(Parameter.DownMixParameter.AsSpan());
             }
 
-            Parameter.DownMixParameter.ToSpan().CopyTo(DownMixCoefficients.AsSpan());
+            Parameter.DownMixParameter.AsSpan().CopyTo(DownMixCoefficients.AsSpan());
 
             errorInfo = new BehaviourParameter.ErrorInfo();
             outStatus = new SinkOutStatus();

Ryujinx.Audio/Renderer/Server/Splitter/SplitterDestination.cs

@@ -172,7 +172,7 @@ namespace Ryujinx.Audio.Renderer.Server.Splitter
         {
             unsafe
             {
-                fixed (SplitterDestination *nextPtr = &next)
+                fixed (SplitterDestination* nextPtr = &next)
                 {
                     _next = nextPtr;
                 }

Ryujinx.Audio/Renderer/Server/StateUpdater.cs

@@ -21,7 +21,7 @@ namespace Ryujinx.Audio.Renderer.Server
     public class StateUpdater
     {
         private readonly ReadOnlyMemory<byte> _inputOrigin;
-        private ReadOnlyMemory <byte> _outputOrigin;
+        private ReadOnlyMemory<byte> _outputOrigin;
         private ReadOnlyMemory<byte> _input;
 
         private Memory<byte> _output;
@@ -119,7 +119,7 @@ namespace Ryujinx.Audio.Renderer.Server
                 ref VoiceChannelResource resource = ref context.GetChannelResource(i);
 
                 resource.Id = parameter.Id;
-                parameter.Mix.ToSpan().CopyTo(resource.Mix.ToSpan());
+                parameter.Mix.AsSpan().CopyTo(resource.Mix.AsSpan());
                 resource.IsUsed = parameter.IsUsed;
             }
 
@@ -207,7 +207,7 @@ namespace Ryujinx.Audio.Renderer.Server
             return ResultCode.Success;
         }
 
-        private static void ResetEffect<T>(ref BaseEffect effect, ref T parameter, PoolMapper mapper) where T: unmanaged, IEffectInParameter
+        private static void ResetEffect<T>(ref BaseEffect effect, ref T parameter, PoolMapper mapper) where T : unmanaged, IEffectInParameter
         {
             effect.ForceUnmapBuffers(mapper);
 
@@ -587,7 +587,7 @@ namespace Ryujinx.Audio.Renderer.Server
         {
             ref BehaviourErrorInfoOutStatus outStatus = ref SpanIOHelper.GetWriteRef<BehaviourErrorInfoOutStatus>(ref _output)[0];
 
-            _behaviourContext.CopyErrorInfo(outStatus.ErrorInfos.ToSpan(), out outStatus.ErrorInfosCount);
+            _behaviourContext.CopyErrorInfo(outStatus.ErrorInfos.AsSpan(), out outStatus.ErrorInfosCount);
 
             OutputHeader.BehaviourSize = (uint)Unsafe.SizeOf<BehaviourErrorInfoOutStatus>();
             OutputHeader.TotalSize += OutputHeader.BehaviourSize;

Ryujinx.Audio/Renderer/Server/Voice/VoiceChannelResource.cs

@@ -34,7 +34,7 @@ namespace Ryujinx.Audio.Renderer.Server.Voice
 
         public void UpdateState()
         {
-            Mix.ToSpan().CopyTo(PreviousMix.ToSpan());
+            Mix.AsSpan().CopyTo(PreviousMix.AsSpan());
         }
     }
 }

Ryujinx.Audio/Renderer/Server/Voice/VoiceState.cs

@@ -200,9 +200,9 @@ namespace Ryujinx.Audio.Renderer.Server.Voice
             SampleFormat = SampleFormat.Invalid;
             ChannelsCount = 0;
             Pitch = 0.0f;
-            Volume= 0.0f;
+            Volume = 0.0f;
             PreviousVolume = 0.0f;
-            BiquadFilters.ToSpan().Fill(new BiquadFilterParameter());
+            BiquadFilters.AsSpan().Fill(new BiquadFilterParameter());
             WaveBuffersCount = 0;
             WaveBuffersIndex = 0;
             MixId = Constants.UnusedMixId;
@@ -288,7 +288,7 @@ namespace Ryujinx.Audio.Renderer.Server.Voice
             ChannelsCount = parameter.ChannelCount;
             Pitch = parameter.Pitch;
             Volume = parameter.Volume;
-            parameter.BiquadFilters.ToSpan().CopyTo(BiquadFilters.ToSpan());
+            parameter.BiquadFilters.AsSpan().CopyTo(BiquadFilters.AsSpan());
             WaveBuffersCount = parameter.WaveBuffersCount;
             WaveBuffersIndex = parameter.WaveBuffersIndex;
 
@@ -308,7 +308,7 @@ namespace Ryujinx.Audio.Renderer.Server.Voice
                 SplitterId = Constants.UnusedSplitterId;
             }
 
-            parameter.ChannelResourceIds.ToSpan().CopyTo(ChannelResourceIds.ToSpan());
+            parameter.ChannelResourceIds.AsSpan().CopyTo(ChannelResourceIds.AsSpan());
 
             DecodingBehaviour behaviour = DecodingBehaviour.Default;
 
@@ -638,7 +638,7 @@ namespace Ryujinx.Audio.Renderer.Server.Voice
 
                         voiceUpdateState.Offset = 0;
                         voiceUpdateState.PlayedSampleCount = 0;
-                        voiceUpdateState.Pitch.ToSpan().Fill(0);
+                        voiceUpdateState.Pitch.AsSpan().Fill(0);
                         voiceUpdateState.Fraction = 0;
                         voiceUpdateState.LoopContext = new Dsp.State.AdpcmLoopContext();
                     }
@@ -650,7 +650,7 @@ namespace Ryujinx.Audio.Renderer.Server.Voice
 
                 case Types.PlayState.Stopped:
                 case Types.PlayState.Paused:
-                    foreach (ref WaveBuffer wavebuffer in WaveBuffers.ToSpan())
+                    foreach (ref WaveBuffer wavebuffer in WaveBuffers.AsSpan())
                     {
                         wavebuffer.BufferAddressInfo.GetReference(true);
                         wavebuffer.ContextAddressInfo.GetReference(true);

Ryujinx.Audio/Renderer/Utils/AudioProcessorMemoryManager.cs

@@ -1,7 +1,6 @@
 using System.Runtime.CompilerServices;
-
-using DspAddress = System.UInt64;
 using CpuAddress = System.UInt64;
+using DspAddress = System.UInt64;
 
 namespace Ryujinx.Audio.Renderer.Utils
 {

Ryujinx.Ava/App.axaml.cs

@@ -55,7 +55,6 @@ namespace Ryujinx.Ava
                 if (ApplicationLifetime is IClassicDesktopStyleApplicationLifetime desktop)
                 {
                     var result = await ContentDialogHelper.CreateConfirmationDialog(
-                        (desktop.MainWindow as MainWindow).SettingsWindow,
                         LocaleManager.Instance["DialogThemeRestartMessage"],
                         LocaleManager.Instance["DialogThemeRestartSubMessage"],
                         LocaleManager.Instance["InputDialogYes"],

Ryujinx.Ava/AppHost.cs

@@ -1,5 +1,6 @@
 using ARMeilleure.Translation;
 using ARMeilleure.Translation.PTC;
+using Avalonia;
 using Avalonia.Input;
 using Avalonia.Threading;
 using LibHac.Tools.FsSystem;
@@ -11,8 +12,10 @@ using Ryujinx.Audio.Integration;
 using Ryujinx.Ava.Common;
 using Ryujinx.Ava.Common.Locale;
 using Ryujinx.Ava.Input;
+using Ryujinx.Ava.Ui.Backend.Vulkan;
 using Ryujinx.Ava.Ui.Controls;
 using Ryujinx.Ava.Ui.Models;
+using Ryujinx.Ava.Ui.Vulkan;
 using Ryujinx.Ava.Ui.Windows;
 using Ryujinx.Common;
 using Ryujinx.Common.Configuration;
@@ -22,6 +25,7 @@ using Ryujinx.Graphics.GAL;
 using Ryujinx.Graphics.GAL.Multithreading;
 using Ryujinx.Graphics.Gpu;
 using Ryujinx.Graphics.OpenGL;
+using Ryujinx.Graphics.Vulkan;
 using Ryujinx.HLE.FileSystem;
 using Ryujinx.HLE.HOS;
 using Ryujinx.HLE.HOS.Services.Account.Acc;
@@ -53,6 +57,7 @@ namespace Ryujinx.Ava
     internal class AppHost
     {
         private const int CursorHideIdleTime = 8; // Hide Cursor seconds
+        private const float MaxResolutionScale = 4.0f; // Max resolution hotkeys can scale to before wrapping.
 
         private static readonly Cursor InvisibleCursor = new Cursor(StandardCursorType.None);
 
@@ -330,6 +335,8 @@ namespace Ryujinx.Ava
                 return;
             }
 
+            AvaloniaLocator.Current.GetService<VulkanPlatformInterface>()?.MainSurface.Display.ChangeVSyncMode(true);
+
             _isStopped = true;
             _isActive = false;
         }
@@ -365,6 +372,7 @@ namespace Ryujinx.Ava
             ConfigurationState.Instance.System.IgnoreMissingServices.Event -= UpdateIgnoreMissingServicesState;
             ConfigurationState.Instance.Graphics.AspectRatio.Event -= UpdateAspectRatioState;
             ConfigurationState.Instance.System.EnableDockedMode.Event -= UpdateDockedModeState;
+            ConfigurationState.Instance.System.AudioVolume.Event -= UpdateAudioVolumeState;
 
             _gpuCancellationTokenSource.Cancel();
             _gpuCancellationTokenSource.Dispose();
@@ -417,10 +425,12 @@ namespace Ryujinx.Ava
                 {
                     if (userError == UserError.NoFirmware)
                     {
-                        string message = string.Format(LocaleManager.Instance["DialogFirmwareInstallEmbeddedMessage"], firmwareVersion.VersionString);
+                        string message = string.Format(LocaleManager.Instance["DialogFirmwareInstallEmbeddedMessage"],
+                            firmwareVersion.VersionString);
 
-                        UserResult result = await ContentDialogHelper.CreateConfirmationDialog(_parent,
-                            LocaleManager.Instance["DialogFirmwareNoFirmwareInstalledMessage"], message, LocaleManager.Instance["InputDialogYes"], LocaleManager.Instance["InputDialogNo"], "");
+                        UserResult result = await ContentDialogHelper.CreateConfirmationDialog(
+                            LocaleManager.Instance["DialogFirmwareNoFirmwareInstalledMessage"], message,
+                            LocaleManager.Instance["InputDialogYes"], LocaleManager.Instance["InputDialogNo"], "");
 
                         if (result != UserResult.Yes)
                         {
@@ -450,7 +460,7 @@ namespace Ryujinx.Ava
 
                         string message = string.Format(LocaleManager.Instance["DialogFirmwareInstallEmbeddedSuccessMessage"], firmwareVersion.VersionString);
 
-                        await ContentDialogHelper.CreateInfoDialog(_parent,
+                        await ContentDialogHelper.CreateInfoDialog(
                             string.Format(LocaleManager.Instance["DialogFirmwareInstalledMessage"], firmwareVersion.VersionString),
                             message,
                             LocaleManager.Instance["InputDialogOk"],
@@ -584,7 +594,24 @@ namespace Ryujinx.Ava
         {
             VirtualFileSystem.ReloadKeySet();
 
-            IRenderer renderer = new Renderer();
+            IRenderer renderer;
+
+            if (Program.UseVulkan)
+            {
+                var vulkan = AvaloniaLocator.Current.GetService<VulkanPlatformInterface>();
+                
+                renderer = new VulkanRenderer(vulkan.Instance.InternalHandle,
+                    vulkan.MainSurface.Device.InternalHandle,
+                    vulkan.PhysicalDevice.InternalHandle,
+                    vulkan.MainSurface.Device.Queue.InternalHandle,
+                    vulkan.PhysicalDevice.QueueFamilyIndex,
+                    vulkan.MainSurface.Device.Lock);
+            }
+            else
+            {
+                renderer = new OpenGLRenderer();
+            }
+
             IHardwareDeviceDriver deviceDriver = new DummyHardwareDeviceDriver();
 
             BackendThreading threadingMode = ConfigurationState.Instance.Graphics.BackendThreading;
@@ -752,8 +779,11 @@ namespace Ryujinx.Ava
             Width = (int)e.Width;
             Height = (int)e.Height;
 
+            if (!Program.UseVulkan)
+            {
                 SetRendererWindowSize(e);
             }
+        }
 
         private void MainLoop()
         {
@@ -792,10 +822,12 @@ namespace Ryujinx.Ava
 
             _renderer.ScreenCaptured += Renderer_ScreenCaptured;
 
-            (_renderer as Renderer).InitializeBackgroundContext(SPBOpenGLContext.CreateBackgroundContext(Renderer.GameContext));
+            (_renderer as OpenGLRenderer)?.InitializeBackgroundContext(SPBOpenGLContext.CreateBackgroundContext((Renderer as OpenGLRendererControl).GameContext));
 
             Renderer.MakeCurrent();
 
+            AvaloniaLocator.Current.GetService<VulkanPlatformInterface>()?.MainSurface?.Display?.ChangeVSyncMode(Device.EnableDeviceVsync);
+
             Device.Gpu.Renderer.Initialize(_glLogLevel);
 
             Width = (int)Renderer.Bounds.Width;
@@ -811,8 +843,6 @@ namespace Ryujinx.Ava
 
                 Renderer.Start();
 
-                Renderer.QueueRender();
-
                 while (_isActive)
                 {
                     if (Device.WaitFifo())
@@ -853,16 +883,25 @@ namespace Ryujinx.Ava
                 dockedMode += $" ({scale}x)";
             }
 
-            string vendor = _renderer is Renderer renderer ? renderer.GpuVendor : "";
-
             StatusUpdatedEvent?.Invoke(this, new StatusUpdatedEventArgs(
                 Device.EnableDeviceVsync,
                 Device.GetVolume(),
+                Program.UseVulkan ? "Vulkan" : "OpenGL",
                 dockedMode,
                 ConfigurationState.Instance.Graphics.AspectRatio.Value.ToText(),
                 LocaleManager.Instance["Game"] + $": {Device.Statistics.GetGameFrameRate():00.00} FPS ({Device.Statistics.GetGameFrameTime():00.00} ms)",
                 $"FIFO: {Device.Statistics.GetFifoPercent():00.00} %",
-                $"GPU: {vendor}"));
+                $"GPU: {_renderer.GetHardwareInfo().GpuVendor}"));
+
+            if (Program.UseVulkan)
+            {
+                var platformInterface = AvaloniaLocator.Current.GetService<VulkanPlatformInterface>();
+                if (platformInterface.MainSurface.Display.IsSurfaceChanged())
+                {
+                    SetRendererWindowSize(new Size(Width, Height));
+                    return;
+                }
+            }
 
             Renderer.Present(image);
         }
@@ -879,7 +918,7 @@ namespace Ryujinx.Ava
                 }
 
                 _dialogShown = true;
-                shouldExit = await ContentDialogHelper.CreateStopEmulationDialog(_parent);
+                shouldExit = await ContentDialogHelper.CreateStopEmulationDialog();
 
                 _dialogShown = false;
             }
@@ -945,6 +984,9 @@ namespace Ryujinx.Ava
                     {
                         case KeyboardHotkeyState.ToggleVSync:
                             Device.EnableDeviceVsync = !Device.EnableDeviceVsync;
+
+                            AvaloniaLocator.Current.GetService<VulkanPlatformInterface>()?.MainSurface?.Display?.ChangeVSyncMode(Device.EnableDeviceVsync);
+
                             break;
                         case KeyboardHotkeyState.Screenshot:
                             ScreenshotRequested = true;
@@ -974,6 +1016,13 @@ namespace Ryujinx.Ava
 
                             _parent.ViewModel.Volume = Device.GetVolume();
                             break;
+                        case KeyboardHotkeyState.ResScaleUp:
+                            GraphicsConfig.ResScale = GraphicsConfig.ResScale % MaxResolutionScale + 1;
+                            break;
+                        case KeyboardHotkeyState.ResScaleDown:
+                            GraphicsConfig.ResScale =
+                            (MaxResolutionScale + GraphicsConfig.ResScale - 2) % MaxResolutionScale + 1;
+                            break;
                         case KeyboardHotkeyState.None:
                             (_keyboardInterface as AvaloniaKeyboard).Clear();
                             break;
@@ -1031,6 +1080,14 @@ namespace Ryujinx.Ava
             {
                 state = KeyboardHotkeyState.ToggleMute;
             }
+            else if (_keyboardInterface.IsPressed((Key)ConfigurationState.Instance.Hid.Hotkeys.Value.ResScaleUp))
+            {
+                state = KeyboardHotkeyState.ResScaleUp;
+            }
+            else if (_keyboardInterface.IsPressed((Key)ConfigurationState.Instance.Hid.Hotkeys.Value.ResScaleDown))
+            {
+                state = KeyboardHotkeyState.ResScaleDown;
+            }
 
             return state;
         }

Ryujinx.Ava/Assets/Locales/de_DE.json

@@ -20,7 +20,7 @@
   "MenuBarOptionsManageUserProfiles": "_Profilverwaltung",
   "MenuBarActions": "_Aktionen",
   "MenuBarOptionsSimulateWakeUpMessage": "Aufwachnachricht",
-  "MenuBarActionsScanAmiibo": "Virtualisiere Amiibo",
+  "MenuBarActionsScanAmiibo": "Scanne ein Amiibo",
   "MenuBarTools": "_Werkzeuge",
   "MenuBarToolsInstallFirmware": "Firmware installieren",
   "MenuBarFileToolsInstallFirmwareFromFile": "Installiere Firmware von einer XCI oder einer ZIP Datei",
@@ -52,8 +52,8 @@
   "GameListContextMenuOpenModsDirectory": "Mod-Verzeichnis öffnen",
   "GameListContextMenuOpenModsDirectoryToolTip": "Öffnet das Verzeichnis, welches Mods für die Spiele beinhaltet",
   "GameListContextMenuCacheManagement": "Cache Verwaltung",
-  "GameListContextMenuCacheManagementPurgePptc": "PPTC Cache löschen",
-  "GameListContextMenuCacheManagementPurgePptcToolTip": "Löscht den PPTC Cache der Anwendung",
+  "GameListContextMenuCacheManagementPurgePptc": "PPTC als ungültig markieren",
+  "GameListContextMenuCacheManagementPurgePptcToolTip": "Markiert den PPTC als ungültig, sodass dieser beim nächsten Spielstart neu erstellt wird",
   "GameListContextMenuCacheManagementPurgeShaderCache": "Shader Cache löschen",
   "GameListContextMenuCacheManagementPurgeShaderCacheToolTip": "Löscht den Shader Cache der Anwendung",
   "GameListContextMenuCacheManagementOpenPptcDirectory": "PPTC Verzeichnis öffnen",
@@ -109,21 +109,21 @@
   "SettingsTabSystemSystemLanguageTraditionalChinese": "Traditionelles Chinesisch",
   "SettingsTabSystemSystemTimeZone": "System Zeitzone:",
   "SettingsTabSystemSystemTime": "System Zeit:",
-  "SettingsTabSystemEnableVsync": "Aktiviere VSync",
-  "SettingsTabSystemEnablePptc": "Aktiviere PPTC Cache (Profiled presistent translastion cache)",
-  "SettingsTabSystemEnableFsIntegrityChecks": "Aktiviere die FS Integritätsprüfung",
+  "SettingsTabSystemEnableVsync": "VSync",
+  "SettingsTabSystemEnablePptc": "PPTC Cache (Profiled Persistent Translation Cache)",
+  "SettingsTabSystemEnableFsIntegrityChecks": "FS Integritätsprüfung",
   "SettingsTabSystemAudioBackend": "Audio-Backend:",
   "SettingsTabSystemAudioBackendDummy": "Dummy",
   "SettingsTabSystemAudioBackendOpenAL": "OpenAL",
   "SettingsTabSystemAudioBackendSoundIO": "SoundIO",
   "SettingsTabSystemAudioBackendSDL2": "SDL2",
   "SettingsTabSystemHacks": "Hacks",
-  "SettingsTabSystemHacksNote": " - Können Fehler verursachen",
+  "SettingsTabSystemHacksNote": " (Kann Fehler verursachen)",
   "SettingsTabSystemExpandDramSize": "Erweitere DRAM Größe auf 6GB",
   "SettingsTabSystemIgnoreMissingServices": "Ignoriere fehlende Dienste",
   "SettingsTabGraphics": "Grafik",
-  "SettingsTabGraphicsEnhancements": "Verbesserungen",
-  "SettingsTabGraphicsEnableShaderCache": "Aktiviere den Shader Cache",
+  "SettingsTabGraphicsAPI": "Grafik-API",
+  "SettingsTabGraphicsEnableShaderCache": "Shader Cache",
   "SettingsTabGraphicsAnisotropicFiltering": "Anisotrope Filterung:",
   "SettingsTabGraphicsAnisotropicFilteringAuto": "Auto",
   "SettingsTabGraphicsAnisotropicFiltering2x": "2x",
@@ -185,9 +185,9 @@
   "ControllerSettingsControllerType": "Controller Typ",
   "ControllerSettingsControllerTypeHandheld": "Handheld",
   "ControllerSettingsControllerTypeProController": "Pro Controller",
-  "ControllerSettingsControllerTypeJoyConPair": "JoyCon Paar",
-  "ControllerSettingsControllerTypeJoyConLeft": "JoyCon Links",
-  "ControllerSettingsControllerTypeJoyConRight": "JoyCon Rechts",
+  "ControllerSettingsControllerTypeJoyConPair": "Joy-Con Paar",
+  "ControllerSettingsControllerTypeJoyConLeft": "Joy-Con Links",
+  "ControllerSettingsControllerTypeJoyConRight": "Joy-Con Rechts",
   "ControllerSettingsProfile": "Profil",
   "ControllerSettingsProfileDefault": "Default",
   "ControllerSettingsLoad": "Laden",
@@ -212,8 +212,8 @@
   "ControllerSettingsLStickLeft": "Links",
   "ControllerSettingsLStickRight": "Rechts",
   "ControllerSettingsLStickStick": "Analogstick",
-  "ControllerSettingsLStickInvertXAxis": "Invertiert X-Achse",
-  "ControllerSettingsLStickInvertYAxis": "Invertiert Y-Achse",
+  "ControllerSettingsLStickInvertXAxis": "X-Achse invertieren",
+  "ControllerSettingsLStickInvertYAxis": "Y-Achse invertieren",
   "ControllerSettingsLStickDeadzone": "Deadzone:",
   "ControllerSettingsRStick": "Rechter Analogstick",
   "ControllerSettingsRStickButton": "R3",
@@ -222,8 +222,8 @@
   "ControllerSettingsRStickLeft": "Links",
   "ControllerSettingsRStickRight": "Rechts",
   "ControllerSettingsRStickStick": "Analogstick",
-  "ControllerSettingsRStickInvertXAxis": "Invertiert X-Achse",
-  "ControllerSettingsRStickInvertYAxis": "Invertiert Y-Achse",
+  "ControllerSettingsRStickInvertXAxis": "X-Achse invertieren",
+  "ControllerSettingsRStickInvertYAxis": "Y-Achse invertieren",
   "ControllerSettingsRStickDeadzone": "Deadzone:",
   "ControllerSettingsTriggersLeft": "Linker Trigger",
   "ControllerSettingsTriggersRight": "Rechter Trigger",
@@ -246,7 +246,7 @@
   "ControllerSettingsMotionUseCemuhookCompatibleMotion": "CemuHook kompatible Bewegungssteuerung",
   "ControllerSettingsMotionControllerSlot": "Controller Slot:",
   "ControllerSettingsMotionMirrorInput": "Spiegele Eingabe",
-  "ControllerSettingsMotionRightJoyConSlot": "Rechter JoyCon Slot:",
+  "ControllerSettingsMotionRightJoyConSlot": "Rechter Joy-Con Slot:",
   "ControllerSettingsMotionServerHost": "Server Host:",
   "ControllerSettingsMotionGyroSensitivity": "Gyro Empfindlichkeit:",
   "ControllerSettingsMotionGyroDeadzone": "Gyro Deadzone:",
@@ -256,8 +256,8 @@
   "UserProfilesSaveProfileName": "Profilname speichern",
   "UserProfilesChangeProfileImage": "Profilbild ändern",
   "UserProfilesAvailableUserProfiles": "Verfügbare Profile:",
-  "UserProfilesAddNewProfile": "Neues Profil hinzufügen",
-  "UserProfilesDeleteSelectedProfile": "Ausgewähltes Profil löschen",
+  "UserProfilesAddNewProfile": "Neues Profil",
+  "UserProfilesDeleteSelectedProfile": "Profil löschen",
   "UserProfilesClose": "Schließen",
   "ProfileImageSelectionTitle": "Auswahl des Profilbildes",
   "ProfileImageSelectionHeader": "Wähle ein Profilbild aus",
@@ -281,17 +281,17 @@
   "MenuBarFileToolsHideUi": "Verstecke UI",
   "GameListContextMenuToggleFavorite": "Als Favoriten hinzufügen/entfernen",
   "GameListContextMenuToggleFavoriteToolTip": "Aktiviert den Favoriten-Status des Spiels",
-  "SettingsTabGeneralTheme": "Thema",
-  "SettingsTabGeneralThemeCustomTheme": "Verzeichnis für benutzerdefiniertes Thema",
+  "SettingsTabGeneralTheme": "Design",
+  "SettingsTabGeneralThemeCustomTheme": "Pfad für das benutzerdefinierte Design",
   "SettingsTabGeneralThemeBaseStyle": "Farbschema",
   "SettingsTabGeneralThemeBaseStyleDark": "Dunkel",
   "SettingsTabGeneralThemeBaseStyleLight": "Hell",
-  "SettingsTabGeneralThemeEnableCustomTheme": "Benutzerdefiniertes Thema",
+  "SettingsTabGeneralThemeEnableCustomTheme": "Design für die Emulator-Benutzeroberfläche",
   "ButtonBrowse": "Durchsuchen",
   "ControllerSettingsConfigureGeneral": "Konfigurieren",
   "ControllerSettingsRumble": "Vibration",
-  "ControllerSettingsRumbleStrongMultiplier": "Starke Vibration - Multiplikator",
-  "ControllerSettingsRumbleWeakMultiplier": "Schwache Vibration - Multiplikator",
+  "ControllerSettingsRumbleStrongMultiplier": "Starker Vibrations-Multiplikator",
+  "ControllerSettingsRumbleWeakMultiplier": "Schwacher Vibrations-Multiplikator",
   "DialogMessageSaveNotAvailableMessage": "Es existieren keine Speicherdaten für {0} [{1:x16}]",
   "DialogMessageSaveNotAvailableCreateSaveMessage": "Soll Ryujinx Speicherdaten für dieses Spiel erstellen?",
   "DialogConfirmationTitle": "Ryujinx - Bestätigung",
@@ -328,7 +328,7 @@
   "DialogUpdaterDirtyBuildMessage": "Inoffizielle Versionen von Ryujinx können nicht aktualisiert werden",
   "DialogUpdaterDirtyBuildSubMessage": "Lade Ryujinx bitte von hier herunter, um eine unterstützte Version zu erhalten: https://ryujinx.org/",
   "DialogRestartRequiredMessage": "Neustart erforderlich",
-  "DialogThemeRestartMessage": "Das Thema wurde gespeichert. Ein Neustart ist erforderlich, um das Thema anzuwenden.",
+  "DialogThemeRestartMessage": "Das Design wurde gespeichert. Ein Neustart ist erforderlich, um das Design anzuwenden.",
   "DialogThemeRestartSubMessage": "Jetzt neu starten?",
   "DialogFirmwareInstallEmbeddedMessage": "Die in diesem Spiel enthaltene Firmware installieren? (Firmware {0})",
   "DialogFirmwareInstallEmbeddedSuccessMessage": "Es wurde keine installierte Firmware gefunden, aber Ryujinx konnte die Firmware {0} aus dem bereitgestellten Spiel installieren.\nRyujinx wird nun gestartet.",
@@ -350,7 +350,7 @@
   "DialogProfileDeleteProfileTitle": "Profil löschen",
   "DialogProfileDeleteProfileMessage": "Diese Aktion kann nicht rückgängig gemacht werden. Wirklich fortfahren?",
   "DialogWarning": "Warnung",
-  "DialogPPTCDeletionMessage": "Du bist dabei, den PPTC Cache zu löschen für :\n\n{0}\n\nWirklich fortfahren?",
+  "DialogPPTCDeletionMessage": "Du bist dabei den PPTC für das folgende Spiel als ungültig zu markieren:\n\n{0}\n\nWirklich fortfahren?",
   "DialogPPTCDeletionErrorMessage": "Fehler bei der Löschung des PPTC Caches bei {0}: {1}",
   "DialogShaderDeletionMessage": "Du bist dabei, den Shader Cache zu löschen für :\n\n{0}\n\nWirklich fortfahren?",
   "DialogShaderDeletionErrorMessage": "Es ist ein Fehler bei der Löschung des Shader Caches bei {0}: {1} aufgetreten",
@@ -379,7 +379,7 @@
   "DialogSettingsBackendThreadingWarningTitle": "Warnung - Render Threading",
   "DialogSettingsBackendThreadingWarningMessage": "Ryujinx muss muss neu gestartet werden, damit die Änderungen wirksam werden. Abhängig von dem Betriebssystem muss möglicherweise das Multithreading des Treibers manuell deaktiviert werden, wenn Ryujinx verwendet wird.",
   "SettingsTabGraphicsFeaturesOptions": "Erweiterungen",
-  "SettingsTabGraphicsBackendMultithreading": "Grafikbackend Multithreading:",
+  "SettingsTabGraphicsBackendMultithreading": "Grafik-Backend Multithreading:",
   "CommonAuto": "Auto",
   "CommonOff": "Aus",
   "CommonOn": "An",
@@ -418,50 +418,50 @@
   "OrderDescending": "Absteigend",
   "SettingsTabGraphicsFeatures": "Erweiterungen",
   "ErrorWindowTitle": "Fehler-Fenster",
-  "ToggleDiscordTooltip": "Aktiviert/Deaktiviert Discord Rich Presence",
+  "ToggleDiscordTooltip": "Zeige momentanes Spiel auf Discord",
   "AddGameDirBoxTooltip": "Gibt das Spielverzeichnis an, das der Liste hinzuzufügt wird",
   "AddGameDirTooltip": "Fügt ein neues Spielverzeichnis hinzu",
   "RemoveGameDirTooltip": "Entfernt das ausgewähltes Spielverzeichnis",
-  "CustomThemeCheckTooltip": "Aktiviert/Deaktiviert die benutzerdefinierten Thema in dem GUI",
-  "CustomThemePathTooltip": " Gibt das Verzeichnis zum benutzerdefinierten GUI Thema an",
-  "CustomThemeBrowseTooltip": "Ermöglicht die Suche nach benutzerdefinierten GUI Thema",
-  "DockModeToggleTooltip": "Aktiviert/Deaktiviert den Docked Mode",
+  "CustomThemeCheckTooltip": "Verwende ein eigenes Design für die Emulator-Benutzeroberfläche",
+  "CustomThemePathTooltip": "Gibt den Pfad zum Design für die Emulator-Benutzeroberfläche an",
+  "CustomThemeBrowseTooltip": "Ermöglicht die Suche nach einem benutzerdefinierten Design für die Emulator-Benutzeroberfläche",
+  "DockModeToggleTooltip": "Im gedockten Modus verhält sich das emulierte System wie eine Nintendo Switch im TV Modus. Dies verbessert die grafische Qualität der meisten Spiele. Umgekehrt führt die Deaktivierung dazu, dass sich das emulierte System wie eine Nintendo Switch im Handheld Modus verhält, was die Grafikqualität beeinträchtigt.\n\nKonfiguriere das Eingabegerät für Spieler 1, um im Docked Modus zu spielen; konfiguriere das Controllerprofil via der Handheld Option, wenn geplant wird den Handheld Modus zu nutzen.\n\nIm Zweifelsfall AN lassen.",
   "DirectKeyboardTooltip": "Aktiviert/Deaktiviert den \"Direkter Tastaturzugriff (HID) Unterstützung\" (Ermöglicht die Benutzung der Tastaur als Eingabegerät in Spielen)",
   "DirectMouseTooltip": "Aktiviert/Deaktiviert den \"Direkten Mauszugriff (HID) Unterstützung\" (Ermöglicht die Benutzung der Maus als Eingabegerät in Spielen)",
   "RegionTooltip": "Ändert die Systemregion",
   "LanguageTooltip": "Ändert die Systemsprache",
   "TimezoneTooltip": "Ändert die Systemzeitzone",
   "TimeTooltip": "Ändert die Systemzeit",
-  "VSyncToggleTooltip": "Aktiviert/Deaktiviert die Vertikale Synchronisierung",
-  "PptcToggleTooltip": "Aktiviert/Deaktiviert den PPTC",
-  "FsIntegrityToggleTooltip": "Aktiviert/Deaktiviert die Integritätsprüfung der Spieldateien",
-  "AudioBackendTooltip": "Ändert das Audio-Backend",
-  "MemoryManagerTooltip": "Ändert wie der Gastspeicher abgebildet wird und wie auf ihn zugegriffen wird. Beinflusst die Leistung der emulierten CPU erheblich.",
+  "VSyncToggleTooltip": "Vertikale Synchronisierung der emulierten Konsole. Diese Option ist ein Frame-Limiter für die meisten Spiele; die Deaktivierung kann dazu führen, dass Spiele mit höherer Geschwindigkeit laufen, Ladebildschirme länger benötigen oder hängen bleiben.\n\nKann beim Spielen mit einem frei wählbaren Hotkey ein- und ausgeschaltet werden.\n\nIm Zweifelsfall AN lassen.",
+  "PptcToggleTooltip": "Speichert übersetzte JIT-Funktionen, sodass jene nicht jedes Mal übersetzt werden müssen, wenn das Spiel geladen wird.\n\nVerringert Stottern und die Zeit beim zweiten und den darauffolgenden Startvorgängen eines Spiels erheblich.\n\nIm Zweifelsfall AN lassen.",
+  "FsIntegrityToggleTooltip": "Prüft beim Startvorgang auf beschädigte Dateien und zeigt bei beschädigten Dateien einen Hash-Fehler (Hash Error) im Log an.\n\nDiese Einstellung hat keinen Einfluss auf die Leistung und hilft bei der Fehlersuche.\n\nIm Zweifelsfall AN lassen.",
+  "AudioBackendTooltip": "Ändert das Backend, das zum Rendern von Audio verwendet wird.\n\nSDL2 ist das bevorzugte Audio-Backend, OpenAL und SoundIO sind als Alternativen vorhanden. Dummy wird keinen Audio-Output haben.\n\nIm Zweifelsfall SDL2 auswählen.",
+  "MemoryManagerTooltip": "Ändert wie der Gastspeicher abgebildet wird und wie auf ihn zugegriffen wird. Beinflusst die Leistung der emulierten CPU erheblich.\n\nIm Zweifelsfall Host ungeprüft auswählen.",
   "MemoryManagerSoftwareTooltip": "Verwendung einer Software-Seitentabelle für die Adressumsetzung. Höchste Genauigkeit, aber langsamste Leistung.",
   "MemoryManagerHostTooltip": "Direkte Zuordnung von Speicher im Host-Adressraum. Viel schnellere JIT-Kompilierung und Ausführung.",
   "MemoryManagerUnsafeTooltip": "Direkte Zuordnung des Speichers, aber keine Maskierung der Adresse innerhalb des Gastadressraums vor dem Zugriff. Schneller, aber auf Kosten der Sicherheit. Die Gastanwendung kann von überall in Ryujinx auf den Speicher zugreifen, daher sollte in diesem Modus nur Programme ausgeführt werden denen vertraut wird.",
-  "DRamTooltip": "Erweitert den Speicher des emulierten Systems von 4 GB auf 6 GB",
-  "IgnoreMissingServicesTooltip": "Aktiviert/Deaktiviert die 'Ignoriere fehlende Dienste' Option",
-  "GraphicsBackendThreadingTooltip": "Aktiviert das Grafikbackend Multithreading",
-  "GalThreadingTooltip": "Führt Grafikbackendbefehle auf einem zweiten Thread aus. Ermöglicht Multithreading bei der Shader Kompilierung zur Laufzeit, reduziert Stottern und verbessert die Leistung bei Treibern ohne eigene Multithreading Unterstützung. Geringfügig abweichende Spitzenleistung bei Treibern mit Multithreading. Ryujinx muss möglicherweise neu gestartet werden, um das in den Treiber integrierte Multithreading korrekt zu deaktivieren, oder es muss manuell getan werden, um die beste Leistung zu erzielen.",
-  "ShaderCacheToggleTooltip": "Aktiviert/Deaktiviert Shader Cache",
+  "DRamTooltip": "Erhöht den Arbeitsspeicher des emulierten Systems von 4 GB auf 6 GB.\n\nDies ist nur für Texturenpakete mit höherer Auflösung oder Mods mit 4K-Auflösung nützlich. Diese Option verbessert NICHT die Leistung.\n\nIm Zweifelsfall AUS lassen.",
+  "IgnoreMissingServicesTooltip": "Durch diese Option werden nicht implementierte Dienste der Switch-Firmware ignoriert. Dies kann dabei helfen, Abstürze beim Starten bestimmter Spiele zu umgehen.\n\nIm Zweifelsfall AUS lassen.",
+  "GraphicsBackendThreadingTooltip": "Führt Grafik-Backend Befehle auf einem zweiten Thread aus.\n\nDies beschleunigt die Shader-Kompilierung, reduziert Stottern und verbessert die Leistung auf GPU-Treibern ohne eigene Multithreading-Unterstützung. Geringfügig bessere Leistung bei Treibern mit Multithreading.\n\nIm Zweifelsfall auf AUTO stellen.",
+  "GalThreadingTooltip": "Führt Grafik-Backend Befehle auf einem zweiten Thread aus.\n\nDies Beschleunigt die Shader-Kompilierung, reduziert Stottern und verbessert die Leistung auf GPU-Treibern ohne eigene Multithreading-Unterstützung. Geringfügig bessere Leistung bei Treibern mit Multithreading.\n\nIm Zweifelsfall auf auf AUTO stellen.",
+  "ShaderCacheToggleTooltip": "Speichert einen persistenten Shader Cache, der das Stottern bei nachfolgenden Durchläufen reduziert.\n\nIm Zweifelsfall AN lassen.",
   "ResolutionScaleTooltip": "Wendet die Auflösungsskalierung auf anwendbare Render Ziele",
   "ResolutionScaleEntryTooltip": "Fließkomma Auflösungsskalierung, wie 1,5.\n Bei nicht ganzzahligen Werten ist die Wahrscheinlichkeit größer, dass Probleme entstehen, die auch zum Absturz führen können.",
   "AnisotropyTooltip": "Stufe der Anisotropen Filterung (Auf Auto setzen, um den vom Spiel geforderten Wert zu verwenden)",
   "AspectRatioTooltip": "Auf das Renderer-Fenster angewandtes Seitenverhältnis.",
   "ShaderDumpPathTooltip": "Grafik-Shader-Dump-Pfad",
-  "FileLogTooltip": "Aktiviert/Deaktiviert die Erstellung und Speicherung eines Logs",
-  "StubLogTooltip": "Aktiviert die Ausgabe von Stub-Logs in der Konsole",
-  "InfoLogTooltip": "Aktiviert die Ausgabe von Info-Logs in der Konsole",
-  "WarnLogTooltip": "Aktiviert die Ausgabe von Warn-Logs in der Konsole",
-  "TraceLogTooltip": "Aktiviert die Ausgabe von Trace-Log in der Konsole",
-  "ErrorLogTooltip": "Aktiviert die Ausgabe von Fehler-Logs in der Konsole",
-  "GuestLogTooltip": "Aktiviert die Ausgabe von Gast-Logs in der Konsole",
-  "FileAccessLogTooltip": "Aktiviert die Ausgabe von FS-Zugriff-Logs in der Konsole",
+  "FileLogTooltip": "Speichert die Konsolenausgabe in einer Log-Datei auf der Festplatte. Hat keinen Einfluss auf die Leistung.",
+  "StubLogTooltip": "Ausgabe von Stub-Logs in der Konsole. Hat keinen Einfluss auf die Leistung.",
+  "InfoLogTooltip": "Ausgabe von Info-Logs in der Konsole. Hat keinen Einfluss auf die Leistung.",
+  "WarnLogTooltip": "Ausgabe von Warn-Logs in der Konsole. Hat keinen Einfluss auf die Leistung.",
+  "TraceLogTooltip": "Ausgabe von Trace-Log in der Konsole. Hat keinen Einfluss auf die Leistung.",
+  "ErrorLogTooltip": "Ausgabe von Fehler-Logs in der Konsole. Hat keinen Einfluss auf die Leistung.",
+  "GuestLogTooltip": "Ausgabe von Gast-Logs in der Konsole. Hat keinen Einfluss auf die Leistung.",
+  "FileAccessLogTooltip": "Ausgabe von FS-Zugriff-Logs in der Konsole.",
   "FSAccessLogModeTooltip": "Aktiviert die Ausgabe des FS-Zugriff-Logs in der Konsole. Mögliche Modi sind 0-3",
   "DeveloperOptionTooltip": "Mit Vorsicht verwenden",
   "OpenGlLogLevel": "Erfordert die Aktivierung der entsprechenden Log-Level",
-  "DebugLogTooltip": "Aktiviert das Dokumentierern von Debug-Protokollmeldungen",
+  "DebugLogTooltip": "Ausgabe von Debug-Logs in der Konsole.\n\nVerwende diese Option nur auf ausdrückliche Anweisung von Ryujinx Entwicklern, da sie das Lesen der Protokolle erschwert und die Leistung des Emulators verschlechtert.",
   "LoadApplicationFileTooltip": "Öffnet die Dateiauswahl um Datei zu laden, welche mit der Switch kompatibel ist",
   "LoadApplicationFolderTooltip": "Öffnet die Dateiauswahl um ein Spiel zu laden, welches mit der Switch kompatibel ist",
   "OpenRyujinxFolderTooltip": "Öffnet den Ordner, der das Ryujinx Dateisystem enthält",
@@ -478,8 +478,8 @@
   "AboutRyujinxContributorsButtonHeader": "Alle Mitwirkenden anzeigen",
   "SettingsTabSystemAudioVolume" : "Lautstärke: ",
   "AudioVolumeTooltip": "Ändert die Lautstärke",
-  "SettingsTabSystemEnableInternetAccess": "Aktiviert den Gast-Internet-Zugang",
-  "EnableInternetAccessTooltip": "Aktiviert den Gast-Internet-Zugang. Die Anwendung verhält sich so, als ob die emulierte Switch-Konsole mit dem Internet verbunden wäre. Beachte, dass in einigen Fällen Anwendungen auch bei deaktivierter Option auf das Internet zugreifen können",
+  "SettingsTabSystemEnableInternetAccess": "Gast-Internet-Zugang/LAN Modus",
+  "EnableInternetAccessTooltip": "Erlaubt es der emulierten Anwendung sich mit dem Internet zu verbinden.\n\nSpiele die den LAN-Modus unterstützen, ermöglichen es Ryujinx sich sowohl mit anderen Ryujinx-Systemen, als auch mit offiziellen Nintendo Switch Konsolen zu verbinden. Allerdings nur, wenn diese Option aktiviert ist und die Systeme mit demselben lokalen Netzwerk verbunden sind.\n\nDies erlaubt KEINE Verbindung zu Nintendo-Servern. Kann bei bestimmten Spielen die versuchen sich mit dem Internet zu verbinden zum Absturz führen.\n\nIm Zweifelsfall AUS lassen",
   "GameListContextMenuManageCheatToolTip": "Öffnet den Cheat-Manager",
   "GameListContextMenuManageCheat": "Cheats verwalten",
   "ControllerSettingsStickRange": "Bereich:",
@@ -491,11 +491,11 @@
   "SettingsTabNetworkConnection": "Netwerkverbindung",
   "SettingsTabCpuCache": "CPU-Cache",
   "SettingsTabCpuMemory": "CPU-Speicher",
-  "DialogUpdaterFlatpakNotSupportedMessage": "Bitte Aktualisiere Ryujinx mit FlatHub",
+  "DialogUpdaterFlatpakNotSupportedMessage": "Bitte aktualisiere Ryujinx mit FlatHub",
   "UpdaterDisabledWarningTitle": "Updater deaktiviert!",
   "GameListContextMenuOpenSdModsDirectory": "Atmosphere-Mod-Verzeichnis öffnen",
-  "GameListContextMenuOpenSdModsDirectoryToolTip": "Öffnet das alternative SD-Karten-Atmosphere-Verzeichnis, das die Mods der Anwendung enthält",
-  "ControllerSettingsRotate90": "Rotiert um 90°",
+  "GameListContextMenuOpenSdModsDirectoryToolTip": "Öffnet das alternative SD-Karten-Atmosphere-Verzeichnis, das die Mods der Anwendung enthält. Dieser Ordner ist nützlich für Mods, die für einen gemoddete Switch erstellt worden sind.",
+  "ControllerSettingsRotate90": "Um 90° rotieren",
   "IconSize": "Cover Größe",
   "IconSizeTooltip": "Ändert die Größe der Spiel-Cover",
   "MenuBarOptionsShowConsole": "Zeige Konsole",
@@ -546,15 +546,47 @@
   "RyujinxUpdater": "Ryujinx - Updater",
   "SettingsTabHotkeys": "Tastatur Hotkeys",
   "SettingsTabHotkeysHotkeys": "Tastatur Hotkeys",
-  "SettingsTabHotkeysToggleVsyncHotkey": "Aktiviert/Deaktiviert VSync:",
+  "SettingsTabHotkeysToggleVsyncHotkey": "VSync:",
   "SettingsTabHotkeysScreenshotHotkey": "Screenshot:",
   "SettingsTabHotkeysShowUiHotkey": "Zeige UI:",
   "SettingsTabHotkeysPauseHotkey": "Pausieren:",
   "SettingsTabHotkeysToggleMuteHotkey": "Stummschalten:",
   "ControllerMotionTitle": "Bewegungssteuerung - Einstellungen",
   "ControllerRumbleTitle": "Vibration - Einstellungen",
-  "SettingsSelectThemeFileDialogTitle" : "Wähle ein benutzerdefiniertes Thema",
-  "SettingsXamlThemeFile" : "Xaml Thema-Datei",
-  "SettingsTabGraphicsBackend" : "Grafik-Backend",
-  "GraphicsBackendTooltip" : "Ändert das Grafik-Backend"
+  "SettingsSelectThemeFileDialogTitle" : "Wähle ein Design für die Emulator-Benutzeroberfläche",
+  "SettingsXamlThemeFile" : "Xaml Design-Datei",
+  "AvatarWindowTitle": "Profile verwalten - Avatar",
+  "Amiibo": "Amiibo",
+  "Unknown": "Unbekannt",
+  "Usage": "Nutzung",
+  "Writable": "Beschreibbar",
+  "SelectDlcDialogTitle": "DLC-Dateien auswählen",
+  "SelectUpdateDialogTitle": "Update-Datei auswählen",
+  "UserProfileWindowTitle": "Benutzerprofile verwalten",
+  "CheatWindowTitle": "Spiel-Cheats verwalten",
+  "DlcWindowTitle": "Spiel-DLC verwalten",
+  "UpdateWindowTitle": "Spiel-Updates verwalten",
+  "CheatWindowHeading": "Cheats verfügbar für {0} [{1}]",
+  "DlcWindowHeading": "DLC verfügbar für {0} [{1}]",
+  "UserProfilesEditProfile": "Profil bearbeiten",
+  "Cancel": "Abbrechen",
+  "Save": "Speichern",
+  "Discard": "Verwerfen",
+  "UserProfilesSetProfileImage": "Profilbild einrichten",
+  "UserProfileEmptyNameError": "Name ist erforderlich",
+  "UserProfileNoImageError":  "Bitte ein Profilbild auswählen",
+  "GameUpdateWindowHeading": "Update verfügbar für {0} [{1}]",
+  "SettingsTabHotkeysResScaleUpHotkey": "Auflösung erhöhen:",
+  "SettingsTabHotkeysResScaleDownHotkey": "Auflösung verringern:",
+  "UserProfilesName": "Name:",
+  "UserProfilesUserId" : "Benutzer Id:",
+  "SettingsTabGraphicsBackend": "Grafik-Backend:",
+  "SettingsTabGraphicsBackendTooltip": "Verwendendetes Grafik-Backend",
+  "SettingsEnableTextureRecompression": "Textur-Rekompression",
+  "SettingsEnableTextureRecompressionTooltip": "Komprimiert bestimmte Texturen, um den VRAM-Verbrauch zu reduzieren.\n\nEmpfohlen für die Verwendung von GPUs, die weniger als 4 GB VRAM haben.\n\nIm Zweifelsfall AUS lassen",
+  "SettingsTabGraphicsPreferredGpu": "Bevorzugte GPU:",
+  "SettingsTabGraphicsPreferredGpuTooltip": "Wähle die Grafikkarte aus, die mit dem Vulkan Grafik-Backend verwendet werden soll.\n\nDies hat keinen Einfluss auf die GPU die OpenGL verwendet.\n\nIm Zweifelsfall die als \"dGPU\" gekennzeichnete GPU auswählen. Diese Einstellung unberührt lassen, wenn keine zur Auswahl steht.",
+  "SettingsAppRequiredRestartMessage": "Ein Neustart von Ryujinx ist erforderlich",
+  "SettingsGpuBackendRestartMessage": "Das Grafik-Backend oder die Grafikkarteneinstellungen wurden geändert. Ein Neustart ist erforderlich um diese Einstellungen anzuwenden.",
+  "SettingsGpuBackendRestartSubMessage": "Ryujinx jetzt neu starten?"
 }

Ryujinx.Ava/Assets/Locales/el_GR.json

@@ -118,11 +118,11 @@
   "SettingsTabSystemAudioBackendSoundIO": "SoundIO",
   "SettingsTabSystemAudioBackendSDL2": "SDL2",
   "SettingsTabSystemHacks": "Μικροδιορθώσεις",
-  "SettingsTabSystemHacksNote": " - Μπορεί να προκαλέσουν αστάθεια",
+  "SettingsTabSystemHacksNote": " (Μπορεί να προκαλέσουν αστάθεια)",
   "SettingsTabSystemExpandDramSize": "Επέκταση μεγέθους DRAM στα 6GB",
   "SettingsTabSystemIgnoreMissingServices": "Αγνόηση υπηρεσιών που λείπουν",
   "SettingsTabGraphics": "Γραφικά",
-  "SettingsTabGraphicsEnhancements": "Βελτιώσεις",
+  "SettingsTabGraphicsAPI": "API Γραφικά",
   "SettingsTabGraphicsEnableShaderCache": "Ενεργοποίηση Προσωρινής Μνήμης Shader",
   "SettingsTabGraphicsAnisotropicFiltering": "Ανισότροπο Φιλτράρισμα:",
   "SettingsTabGraphicsAnisotropicFilteringAuto": "Αυτόματο",
@@ -416,7 +416,7 @@
   "CommonFavorite": "Αγαπημένα",
   "OrderAscending": "Αύξουσα",
   "OrderDescending": "Φθίνουσα",
-  "SettingsTabGraphicsFeatures": "Χαρακτηριστικά",
+  "SettingsTabGraphicsFeatures": "Χαρακτηριστικά & Βελτιώσεις",
   "ErrorWindowTitle": "Παράθυρο σφάλματος",
   "ToggleDiscordTooltip": "Ενεργοποιεί ή απενεργοποιεί την Εμπλουτισμένη Παρουσία σας στο Discord",
   "AddGameDirBoxTooltip": "Εισαγάγετε μία τοποθεσία παιχνιδιών για προσθήκη στη λίστα",