New NVDEC and VIC implementation (#1384)

* Initial NVDEC and VIC implementation

* Update FFmpeg.AutoGen to 4.3.0

* Add nvdec dependencies for Windows

* Unify some VP9 structures

* Rename VP9 structure fields

* Improvements to Video API

* XML docs for Common.Memory

* Remove now unused or redundant overloads from MemoryAccessor

* NVDEC UV surface read/write scalar paths

* Add FIXME comments about hacky things/stuff that will need to be fixed in the future

* Cleaned up VP9 memory allocation

* Remove some debug logs

* Rename some VP9 structs

* Remove unused struct

* No need to compile Ryujinx.Graphics.Host1x with unsafe anymore

* Name AsyncWorkQueue threads to make debugging easier

* Make Vp9PictureInfo a ref struct

* LayoutConverter no longer needs the depth argument (broken by rebase)

* Pooling of VP9 buffers, plus fix a memory leak on VP9

* Really wish VS could rename projects properly...

* Address feedback

* Remove using

* Catch OperationCanceledException

* Add licensing informations

* Add THIRDPARTY.md to release too

Co-authored-by: Thog <me@thog.eu>

Ryujinx.Graphics.Nvdec.Vp9/Common/BitUtils.cs

@@ -0,0 +1,59 @@
+using System;
+using System.Diagnostics;
+using System.Numerics;
+using System.Runtime.CompilerServices;
+
+namespace Ryujinx.Graphics.Nvdec.Vp9.Common
+{
+    internal static class BitUtils
+    {
+        // FIXME: Enable inlining here after AVX2 gather bug is fixed.
+        // [MethodImpl(MethodImplOptions.AggressiveInlining)]
+        public static byte ClipPixel(int val)
+        {
+            return (byte)((val > 255) ? 255 : (val < 0) ? 0 : val);
+        }
+
+        [MethodImpl(MethodImplOptions.AggressiveInlining)]
+        public static ushort ClipPixelHighbd(int val, int bd)
+        {
+            return bd switch
+            {
+                10 => (ushort)Math.Clamp(val, 0, 1023),
+                12 => (ushort)Math.Clamp(val, 0, 4095),
+                _ => (ushort)Math.Clamp(val, 0, 255)
+            };
+        }
+
+        [MethodImpl(MethodImplOptions.AggressiveInlining)]
+        public static int RoundPowerOfTwo(int value, int n)
+        {
+            return (value + (1 << (n - 1))) >> n;
+        }
+
+        [MethodImpl(MethodImplOptions.AggressiveInlining)]
+        public static long RoundPowerOfTwo(long value, int n)
+        {
+            return (value + (1L << (n - 1))) >> n;
+        }
+
+        [MethodImpl(MethodImplOptions.AggressiveInlining)]
+        public static int AlignPowerOfTwo(int value, int n)
+        {
+            return (value + ((1 << n) - 1)) & ~((1 << n) - 1);
+        }
+
+        [MethodImpl(MethodImplOptions.AggressiveInlining)]
+        private static int GetMsb(uint n)
+        {
+            Debug.Assert(n != 0);
+            return 31 ^ BitOperations.LeadingZeroCount(n);
+        }
+
+        [MethodImpl(MethodImplOptions.AggressiveInlining)]
+        public static int GetUnsignedBits(uint numValues)
+        {
+            return numValues > 0 ? GetMsb(numValues) + 1 : 0;
+        }
+    }
+}

Ryujinx.Graphics.Nvdec.Vp9/Common/MemoryAllocator.cs

@@ -0,0 +1,94 @@
+using Ryujinx.Common.Memory;
+using System;
+using System.Runtime.CompilerServices;
+using System.Runtime.InteropServices;
+
+namespace Ryujinx.Graphics.Nvdec.Vp9.Common
+{
+    internal class MemoryAllocator : IDisposable
+    {
+        private const int PoolEntries = 10;
+
+        private struct PoolItem
+        {
+            public IntPtr Pointer;
+            public int Length;
+            public bool InUse;
+        }
+
+        private PoolItem[] _pool = new PoolItem[PoolEntries];
+
+        public ArrayPtr<T> Allocate<T>(int length) where T : unmanaged
+        {
+            int lengthInBytes = Unsafe.SizeOf<T>() * length;
+
+            IntPtr ptr = IntPtr.Zero;
+
+            for (int i = 0; i < PoolEntries; i++)
+            {
+                ref PoolItem item = ref _pool[i];
+
+                if (!item.InUse && item.Length == lengthInBytes)
+                {
+                    item.InUse = true;
+                    ptr = item.Pointer;
+                    break;
+                }
+            }
+
+            if (ptr == IntPtr.Zero)
+            {
+                ptr = Marshal.AllocHGlobal(lengthInBytes);
+
+                for (int i = 0; i < PoolEntries; i++)
+                {
+                    ref PoolItem item = ref _pool[i];
+
+                    if (!item.InUse)
+                    {
+                        item.InUse = true;
+                        if (item.Pointer != IntPtr.Zero)
+                        {
+                            Marshal.FreeHGlobal(item.Pointer);
+                        }
+                        item.Pointer = ptr;
+                        item.Length = lengthInBytes;
+                        break;
+                    }
+                }
+            }
+
+            return new ArrayPtr<T>(ptr, length);
+        }
+
+        public unsafe void Free<T>(ArrayPtr<T> arr) where T : unmanaged
+        {
+            IntPtr ptr = (IntPtr)arr.ToPointer();
+
+            for (int i = 0; i < PoolEntries; i++)
+            {
+                ref PoolItem item = ref _pool[i];
+
+                if (item.Pointer == ptr)
+                {
+                    item.InUse = false;
+                    break;
+                }
+            }
+        }
+
+        public void Dispose()
+        {
+            for (int i = 0; i < PoolEntries; i++)
+            {
+                ref PoolItem item = ref _pool[i];
+
+                if (item.Pointer != IntPtr.Zero)
+                {
+                    Marshal.FreeHGlobal(item.Pointer);
+                    item.Pointer = IntPtr.Zero;
+                }
+            }
+        }
+    }
+}

Ryujinx.Graphics.Nvdec.Vp9/Common/MemoryUtil.cs

@@ -0,0 +1,25 @@
+using Ryujinx.Common.Memory;
+using System;
+using System.Runtime.CompilerServices;
+using System.Runtime.InteropServices;
+
+namespace Ryujinx.Graphics.Nvdec.Vp9.Common
+{
+    internal static class MemoryUtil
+    {
+        public static unsafe void Copy<T>(T* dest, T* source, int length) where T : unmanaged
+        {
+            new Span<T>(source, length).CopyTo(new Span<T>(dest, length));
+        }
+
+        public static void Copy<T>(ref T dest, ref T source) where T : unmanaged
+        {
+            MemoryMarshal.CreateSpan(ref source, 1).CopyTo(MemoryMarshal.CreateSpan(ref dest, 1));
+        }
+
+        public static unsafe void Fill<T>(T* ptr, T value, int length) where T : unmanaged
+        {
+            new Span<T>(ptr, length).Fill(value);
+        }
+    }
+}