Ryujinx/Ryujinx.Core/Gpu/NvGpuEngine3d.cs
gdkchan 34037701c7
NvServices refactoring (#120)
* Initial implementation of NvMap/NvHostCtrl

* More work on NvHostCtrl

* Refactoring of nvservices, move GPU Vmm, make Vmm per-process, refactor most gpu devices, move Gpu to Core, fix CbBind

* Implement GetGpuTime, support CancelSynchronization, fix issue on InsertWaitingMutex, proper double buffering support (again, not working properly for commercial games, only hb)

* Try to fix perf regression reading/writing textures, moved syncpts and events to a UserCtx class, delete global state when the process exits, other minor tweaks

* Remove now unused code, add comment about probably wrong result codes
2018-05-07 15:53:23 -03:00

558 lines
18 KiB
C#

using Ryujinx.Graphics.Gal;
using System;
using System.Collections.Generic;
namespace Ryujinx.Core.Gpu
{
public class NvGpuEngine3d : INvGpuEngine
{
public int[] Registers { get; private set; }
private NvGpu Gpu;
private Dictionary<int, NvGpuMethod> Methods;
private struct ConstBuffer
{
public bool Enabled;
public long Position;
public int Size;
}
private ConstBuffer[][] ConstBuffers;
private HashSet<long> FrameBuffers;
public NvGpuEngine3d(NvGpu Gpu)
{
this.Gpu = Gpu;
Registers = new int[0xe00];
Methods = new Dictionary<int, NvGpuMethod>();
void AddMethod(int Meth, int Count, int Stride, NvGpuMethod Method)
{
while (Count-- > 0)
{
Methods.Add(Meth, Method);
Meth += Stride;
}
}
AddMethod(0x585, 1, 1, VertexEndGl);
AddMethod(0x674, 1, 1, ClearBuffers);
AddMethod(0x6c3, 1, 1, QueryControl);
AddMethod(0x8e4, 16, 1, CbData);
AddMethod(0x904, 5, 8, CbBind);
ConstBuffers = new ConstBuffer[6][];
for (int Index = 0; Index < ConstBuffers.Length; Index++)
{
ConstBuffers[Index] = new ConstBuffer[18];
}
FrameBuffers = new HashSet<long>();
}
public void CallMethod(NvGpuVmm Vmm, NvGpuPBEntry PBEntry)
{
if (Methods.TryGetValue(PBEntry.Method, out NvGpuMethod Method))
{
Method(Vmm, PBEntry);
}
else
{
WriteRegister(PBEntry);
}
}
private void VertexEndGl(NvGpuVmm Vmm, NvGpuPBEntry PBEntry)
{
SetFrameBuffer(Vmm, 0);
long[] Tags = UploadShaders(Vmm);
Gpu.Renderer.BindProgram();
SetAlphaBlending();
UploadTextures(Vmm, Tags);
UploadUniforms(Vmm);
UploadVertexArrays(Vmm);
}
private void ClearBuffers(NvGpuVmm Vmm, NvGpuPBEntry PBEntry)
{
int Arg0 = PBEntry.Arguments[0];
int FbIndex = (Arg0 >> 6) & 0xf;
int Layer = (Arg0 >> 10) & 0x3ff;
GalClearBufferFlags Flags = (GalClearBufferFlags)(Arg0 & 0x3f);
SetFrameBuffer(Vmm, 0);
//TODO: Enable this once the frame buffer problems are fixed.
//Gpu.Renderer.ClearBuffers(Layer, Flags);
}
private void SetFrameBuffer(NvGpuVmm Vmm, int FbIndex)
{
long VA = MakeInt64From2xInt32(NvGpuEngine3dReg.FrameBufferNAddress + FbIndex * 0x10);
long PA = Vmm.GetPhysicalAddress(VA);
FrameBuffers.Add(PA);
int Width = ReadRegister(NvGpuEngine3dReg.FrameBufferNWidth + FbIndex * 0x10);
int Height = ReadRegister(NvGpuEngine3dReg.FrameBufferNHeight + FbIndex * 0x10);
//Note: Using the Width/Height results seems to give incorrect results.
//Maybe the size of all frame buffers is hardcoded to screen size? This seems unlikely.
Gpu.Renderer.CreateFrameBuffer(PA, 1280, 720);
Gpu.Renderer.BindFrameBuffer(PA);
}
private long[] UploadShaders(NvGpuVmm Vmm)
{
long[] Tags = new long[5];
long BasePosition = MakeInt64From2xInt32(NvGpuEngine3dReg.ShaderAddress);
for (int Index = 0; Index < 6; Index++)
{
int Control = ReadRegister(NvGpuEngine3dReg.ShaderNControl + Index * 0x10);
int Offset = ReadRegister(NvGpuEngine3dReg.ShaderNOffset + Index * 0x10);
//Note: Vertex Program (B) is always enabled.
bool Enable = (Control & 1) != 0 || Index == 1;
if (!Enable)
{
continue;
}
long Tag = BasePosition + (uint)Offset;
//TODO: Find a better way to calculate the size.
int Size = 0x20000;
byte[] Code = Vmm.ReadBytes(Tag, Size);
GalShaderType ShaderType = GetTypeFromProgram(Index);
Tags[(int)ShaderType] = Tag;
Gpu.Renderer.CreateShader(Tag, ShaderType, Code);
Gpu.Renderer.BindShader(Tag);
}
int RawSX = ReadRegister(NvGpuEngine3dReg.ViewportScaleX);
int RawSY = ReadRegister(NvGpuEngine3dReg.ViewportScaleY);
float SX = BitConverter.Int32BitsToSingle(RawSX);
float SY = BitConverter.Int32BitsToSingle(RawSY);
float SignX = MathF.Sign(SX);
float SignY = MathF.Sign(SY);
Gpu.Renderer.SetUniform2F(GalConsts.FlipUniformName, SignX, SignY);
return Tags;
}
private static GalShaderType GetTypeFromProgram(int Program)
{
switch (Program)
{
case 0:
case 1: return GalShaderType.Vertex;
case 2: return GalShaderType.TessControl;
case 3: return GalShaderType.TessEvaluation;
case 4: return GalShaderType.Geometry;
case 5: return GalShaderType.Fragment;
}
throw new ArgumentOutOfRangeException(nameof(Program));
}
private void SetAlphaBlending()
{
//TODO: Support independent blend properly.
bool Enable = (ReadRegister(NvGpuEngine3dReg.IBlendNEnable) & 1) != 0;
Gpu.Renderer.SetBlendEnable(Enable);
bool BlendSeparateAlpha = (ReadRegister(NvGpuEngine3dReg.IBlendNSeparateAlpha) & 1) != 0;
GalBlendEquation EquationRgb = (GalBlendEquation)ReadRegister(NvGpuEngine3dReg.IBlendNEquationRgb);
GalBlendFactor FuncSrcRgb = (GalBlendFactor)ReadRegister(NvGpuEngine3dReg.IBlendNFuncSrcRgb);
GalBlendFactor FuncDstRgb = (GalBlendFactor)ReadRegister(NvGpuEngine3dReg.IBlendNFuncDstRgb);
if (BlendSeparateAlpha)
{
GalBlendEquation EquationAlpha = (GalBlendEquation)ReadRegister(NvGpuEngine3dReg.IBlendNEquationAlpha);
GalBlendFactor FuncSrcAlpha = (GalBlendFactor)ReadRegister(NvGpuEngine3dReg.IBlendNFuncSrcAlpha);
GalBlendFactor FuncDstAlpha = (GalBlendFactor)ReadRegister(NvGpuEngine3dReg.IBlendNFuncDstAlpha);
Gpu.Renderer.SetBlendSeparate(
EquationRgb,
EquationAlpha,
FuncSrcRgb,
FuncDstRgb,
FuncSrcAlpha,
FuncDstAlpha);
}
else
{
Gpu.Renderer.SetBlend(EquationRgb, FuncSrcRgb, FuncDstRgb);
}
}
private void UploadTextures(NvGpuVmm Vmm, long[] Tags)
{
long BaseShPosition = MakeInt64From2xInt32(NvGpuEngine3dReg.ShaderAddress);
int TextureCbIndex = ReadRegister(NvGpuEngine3dReg.TextureCbIndex);
//Note: On the emulator renderer, Texture Unit 0 is
//reserved for drawing the frame buffer.
int TexIndex = 1;
for (int Index = 0; Index < Tags.Length; Index++)
{
foreach (ShaderDeclInfo DeclInfo in Gpu.Renderer.GetTextureUsage(Tags[Index]))
{
long Position = ConstBuffers[Index][TextureCbIndex].Position;
UploadTexture(Vmm, Position, TexIndex, DeclInfo.Index);
Gpu.Renderer.SetUniform1(DeclInfo.Name, TexIndex);
TexIndex++;
}
}
}
private void UploadTexture(NvGpuVmm Vmm, long BasePosition, int TexIndex, int HndIndex)
{
long Position = BasePosition + HndIndex * 4;
int TextureHandle = Vmm.ReadInt32(Position);
int TicIndex = (TextureHandle >> 0) & 0xfffff;
int TscIndex = (TextureHandle >> 20) & 0xfff;
long TicPosition = MakeInt64From2xInt32(NvGpuEngine3dReg.TexHeaderPoolOffset);
long TscPosition = MakeInt64From2xInt32(NvGpuEngine3dReg.TexSamplerPoolOffset);
TicPosition += TicIndex * 0x20;
TscPosition += TscIndex * 0x20;
GalTextureSampler Sampler = TextureFactory.MakeSampler(Gpu, Vmm, TscPosition);
long TextureAddress = Vmm.ReadInt64(TicPosition + 4) & 0xffffffffffff;
TextureAddress = Vmm.GetPhysicalAddress(TextureAddress);
if (IsFrameBufferPosition(TextureAddress))
{
//This texture is a frame buffer texture,
//we shouldn't read anything from memory and bind
//the frame buffer texture instead, since we're not
//really writing anything to memory.
Gpu.Renderer.BindFrameBufferTexture(TextureAddress, TexIndex, Sampler);
}
else
{
GalTexture Texture = TextureFactory.MakeTexture(Gpu, Vmm, TicPosition);
Gpu.Renderer.SetTextureAndSampler(TexIndex, Texture, Sampler);
Gpu.Renderer.BindTexture(TexIndex);
}
}
private void UploadUniforms(NvGpuVmm Vmm)
{
long BasePosition = MakeInt64From2xInt32(NvGpuEngine3dReg.ShaderAddress);
for (int Index = 0; Index < 5; Index++)
{
int Control = ReadRegister(NvGpuEngine3dReg.ShaderNControl + (Index + 1) * 0x10);
int Offset = ReadRegister(NvGpuEngine3dReg.ShaderNOffset + (Index + 1) * 0x10);
//Note: Vertex Program (B) is always enabled.
bool Enable = (Control & 1) != 0 || Index == 0;
if (!Enable)
{
continue;
}
for (int Cbuf = 0; Cbuf < ConstBuffers.Length; Cbuf++)
{
ConstBuffer Cb = ConstBuffers[Index][Cbuf];
if (Cb.Enabled)
{
byte[] Data = Vmm.ReadBytes(Cb.Position, (uint)Cb.Size);
Gpu.Renderer.SetConstBuffer(BasePosition + (uint)Offset, Cbuf, Data);
}
}
}
}
private void UploadVertexArrays(NvGpuVmm Vmm)
{
long IndexPosition = MakeInt64From2xInt32(NvGpuEngine3dReg.IndexArrayAddress);
int IndexSize = ReadRegister(NvGpuEngine3dReg.IndexArrayFormat);
int IndexFirst = ReadRegister(NvGpuEngine3dReg.IndexBatchFirst);
int IndexCount = ReadRegister(NvGpuEngine3dReg.IndexBatchCount);
GalIndexFormat IndexFormat = (GalIndexFormat)IndexSize;
IndexSize = 1 << IndexSize;
if (IndexSize > 4)
{
throw new InvalidOperationException();
}
if (IndexSize != 0)
{
int BufferSize = IndexCount * IndexSize;
byte[] Data = Vmm.ReadBytes(IndexPosition, BufferSize);
Gpu.Renderer.SetIndexArray(Data, IndexFormat);
}
List<GalVertexAttrib>[] Attribs = new List<GalVertexAttrib>[32];
for (int Attr = 0; Attr < 16; Attr++)
{
int Packed = ReadRegister(NvGpuEngine3dReg.VertexAttribNFormat + Attr);
int ArrayIndex = Packed & 0x1f;
if (Attribs[ArrayIndex] == null)
{
Attribs[ArrayIndex] = new List<GalVertexAttrib>();
}
Attribs[ArrayIndex].Add(new GalVertexAttrib(
Attr,
((Packed >> 6) & 0x1) != 0,
(Packed >> 7) & 0x3fff,
(GalVertexAttribSize)((Packed >> 21) & 0x3f),
(GalVertexAttribType)((Packed >> 27) & 0x7),
((Packed >> 31) & 0x1) != 0));
}
for (int Index = 0; Index < 32; Index++)
{
int VertexFirst = ReadRegister(NvGpuEngine3dReg.VertexArrayFirst);
int VertexCount = ReadRegister(NvGpuEngine3dReg.VertexArrayCount);
int Control = ReadRegister(NvGpuEngine3dReg.VertexArrayNControl + Index * 4);
bool Enable = (Control & 0x1000) != 0;
long VertexPosition = MakeInt64From2xInt32(NvGpuEngine3dReg.VertexArrayNAddress + Index * 4);
if (!Enable)
{
continue;
}
int Stride = Control & 0xfff;
long Size = 0;
if (IndexCount != 0)
{
Size = GetVertexCountFromIndexBuffer(
Vmm,
IndexPosition,
IndexCount,
IndexSize);
}
else
{
Size = VertexCount;
}
//TODO: Support cases where the Stride is 0.
//In this case, we need to use the size of the attribute.
Size *= Stride;
byte[] Data = Vmm.ReadBytes(VertexPosition, Size);
GalVertexAttrib[] AttribArray = Attribs[Index]?.ToArray() ?? new GalVertexAttrib[0];
Gpu.Renderer.SetVertexArray(Index, Stride, Data, AttribArray);
int PrimCtrl = ReadRegister(NvGpuEngine3dReg.VertexBeginGl);
GalPrimitiveType PrimType = (GalPrimitiveType)(PrimCtrl & 0xffff);
if (IndexCount != 0)
{
Gpu.Renderer.DrawElements(Index, IndexFirst, PrimType);
}
else
{
Gpu.Renderer.DrawArrays(Index, VertexFirst, VertexCount, PrimType);
}
}
}
private int GetVertexCountFromIndexBuffer(
NvGpuVmm Vmm,
long IndexPosition,
int IndexCount,
int IndexSize)
{
int MaxIndex = -1;
if (IndexSize == 2)
{
while (IndexCount -- > 0)
{
ushort Value = Vmm.ReadUInt16(IndexPosition);
IndexPosition += 2;
if (MaxIndex < Value)
{
MaxIndex = Value;
}
}
}
else if (IndexSize == 1)
{
while (IndexCount -- > 0)
{
byte Value = Vmm.ReadByte(IndexPosition++);
if (MaxIndex < Value)
{
MaxIndex = Value;
}
}
}
else if (IndexSize == 4)
{
while (IndexCount -- > 0)
{
uint Value = Vmm.ReadUInt32(IndexPosition);
IndexPosition += 2;
if (MaxIndex < Value)
{
MaxIndex = (int)Value;
}
}
}
else
{
throw new ArgumentOutOfRangeException(nameof(IndexSize));
}
return MaxIndex + 1;
}
private void QueryControl(NvGpuVmm Vmm, NvGpuPBEntry PBEntry)
{
long Position = MakeInt64From2xInt32(NvGpuEngine3dReg.QueryAddress);
int Seq = Registers[(int)NvGpuEngine3dReg.QuerySequence];
int Ctrl = Registers[(int)NvGpuEngine3dReg.QueryControl];
int Mode = Ctrl & 3;
if (Mode == 0)
{
//Write mode.
Vmm.WriteInt32(Position, Seq);
}
WriteRegister(PBEntry);
}
private void CbData(NvGpuVmm Vmm, NvGpuPBEntry PBEntry)
{
long Position = MakeInt64From2xInt32(NvGpuEngine3dReg.ConstBufferAddress);
int Offset = ReadRegister(NvGpuEngine3dReg.ConstBufferOffset);
foreach (int Arg in PBEntry.Arguments)
{
Vmm.WriteInt32(Position + Offset, Arg);
Offset += 4;
}
WriteRegister(NvGpuEngine3dReg.ConstBufferOffset, Offset);
}
private void CbBind(NvGpuVmm Vmm, NvGpuPBEntry PBEntry)
{
int Stage = (PBEntry.Method - 0x904) >> 3;
int Index = PBEntry.Arguments[0];
bool Enabled = (Index & 1) != 0;
Index = (Index >> 4) & 0x1f;
long Position = MakeInt64From2xInt32(NvGpuEngine3dReg.ConstBufferAddress);
ConstBuffers[Stage][Index].Position = Position;
ConstBuffers[Stage][Index].Enabled = Enabled;
ConstBuffers[Stage][Index].Size = ReadRegister(NvGpuEngine3dReg.ConstBufferSize);
}
private long MakeInt64From2xInt32(NvGpuEngine3dReg Reg)
{
return
(long)Registers[(int)Reg + 0] << 32 |
(uint)Registers[(int)Reg + 1];
}
private void WriteRegister(NvGpuPBEntry PBEntry)
{
int ArgsCount = PBEntry.Arguments.Count;
if (ArgsCount > 0)
{
Registers[PBEntry.Method] = PBEntry.Arguments[ArgsCount - 1];
}
}
private int ReadRegister(NvGpuEngine3dReg Reg)
{
return Registers[(int)Reg];
}
private void WriteRegister(NvGpuEngine3dReg Reg, int Value)
{
Registers[(int)Reg] = Value;
}
public bool IsFrameBufferPosition(long Position)
{
return FrameBuffers.Contains(Position);
}
}
}