0
0
Fork 0
This repository has been archived on 2024-10-12. You can view files and clone it, but cannot push or open issues or pull requests.
ryujinx-final/Ryujinx.Graphics/NvGpuFifo.cs
gdkchan d2bb458b51 Improved GPU command lists decoding (#499)
* Better implementation of the DMA pusher, misc fixes

* Remove some debug code

* Correct RGBX8 format

* Add support for linked Texture Sampler Control

* Attempt to fix upside down screen issue
2018-11-17 05:01:31 +01:00

176 lines
No EOL
5 KiB
C#

using Ryujinx.Graphics.Memory;
namespace Ryujinx.Graphics
{
class NvGpuFifo
{
private const int MacrosCount = 0x80;
private const int MacroIndexMask = MacrosCount - 1;
//Note: The size of the macro memory is unknown, we just make
//a guess here and use 256kb as the size. Increase if needed.
private const int MmeWords = 256 * 256;
private NvGpu Gpu;
private NvGpuEngine[] SubChannels;
private struct CachedMacro
{
public int Position { get; private set; }
private bool ExecutionPending;
private int Argument;
private MacroInterpreter Interpreter;
public CachedMacro(NvGpuFifo PFifo, INvGpuEngine Engine, int Position)
{
this.Position = Position;
ExecutionPending = false;
Argument = 0;
Interpreter = new MacroInterpreter(PFifo, Engine);
}
public void StartExecution(int Argument)
{
this.Argument = Argument;
ExecutionPending = true;
}
public void Execute(NvGpuVmm Vmm, int[] Mme)
{
if (ExecutionPending)
{
ExecutionPending = false;
Interpreter?.Execute(Vmm, Mme, Position, Argument);
}
}
public void PushArgument(int Argument)
{
Interpreter?.Fifo.Enqueue(Argument);
}
}
private int CurrMacroPosition;
private int CurrMacroBindIndex;
private CachedMacro[] Macros;
private int[] Mme;
public NvGpuFifo(NvGpu Gpu)
{
this.Gpu = Gpu;
SubChannels = new NvGpuEngine[8];
Macros = new CachedMacro[MacrosCount];
Mme = new int[MmeWords];
}
public void CallMethod(NvGpuVmm Vmm, GpuMethodCall MethCall)
{
if ((NvGpuFifoMeth)MethCall.Method == NvGpuFifoMeth.BindChannel)
{
NvGpuEngine Engine = (NvGpuEngine)MethCall.Argument;
SubChannels[MethCall.SubChannel] = Engine;
}
else
{
switch (SubChannels[MethCall.SubChannel])
{
case NvGpuEngine._2d: Call2dMethod (Vmm, MethCall); break;
case NvGpuEngine._3d: Call3dMethod (Vmm, MethCall); break;
case NvGpuEngine.P2mf: CallP2mfMethod(Vmm, MethCall); break;
case NvGpuEngine.M2mf: CallM2mfMethod(Vmm, MethCall); break;
}
}
}
private void Call2dMethod(NvGpuVmm Vmm, GpuMethodCall MethCall)
{
Gpu.Engine2d.CallMethod(Vmm, MethCall);
}
private void Call3dMethod(NvGpuVmm Vmm, GpuMethodCall MethCall)
{
if (MethCall.Method < 0x80)
{
switch ((NvGpuFifoMeth)MethCall.Method)
{
case NvGpuFifoMeth.SetMacroUploadAddress:
{
CurrMacroPosition = MethCall.Argument;
break;
}
case NvGpuFifoMeth.SendMacroCodeData:
{
Mme[CurrMacroPosition++] = MethCall.Argument;
break;
}
case NvGpuFifoMeth.SetMacroBindingIndex:
{
CurrMacroBindIndex = MethCall.Argument;
break;
}
case NvGpuFifoMeth.BindMacro:
{
int Position = MethCall.Argument;
Macros[CurrMacroBindIndex] = new CachedMacro(this, Gpu.Engine3d, Position);
break;
}
default: CallP2mfMethod(Vmm, MethCall); break;
}
}
else if (MethCall.Method < 0xe00)
{
Gpu.Engine3d.CallMethod(Vmm, MethCall);
}
else
{
int MacroIndex = (MethCall.Method >> 1) & MacroIndexMask;
if ((MethCall.Method & 1) != 0)
{
Macros[MacroIndex].PushArgument(MethCall.Argument);
}
else
{
Macros[MacroIndex].StartExecution(MethCall.Argument);
}
if (MethCall.IsLastCall)
{
Macros[MacroIndex].Execute(Vmm, Mme);
}
}
}
private void CallP2mfMethod(NvGpuVmm Vmm, GpuMethodCall MethCall)
{
Gpu.EngineP2mf.CallMethod(Vmm, MethCall);
}
private void CallM2mfMethod(NvGpuVmm Vmm, GpuMethodCall MethCall)
{
Gpu.EngineM2mf.CallMethod(Vmm, MethCall);
}
}
}