forked from Mirror/Ryujinx
5d69d9103e
* Initial implementation. Still pending better valid-overlap handling, disposed pool, compressed format flush fix. * Very messy backend resource cache. * Oops * Dispose -> Release * Improve Release/Dispose. * More rule refinement. * View compatibility levels as an enum - you can always know if a view is only copy compatible. * General cleanup. Use locking on the resource cache, as it is likely to be used by other threads in future. * Rename resource cache to resource pool. * Address some of the smaller nits. * Fix regression with MK8 lens flare Texture flushes done the old way should trigger memory tracking. * Use TextureCreateInfo as a key. It now implements IEquatable and generates a hashcode based on width/height. * Fix size change for compressed+non-compressed view combos. Before, this could set either the compressed or non compressed texture with a size with the wrong size, depending on which texture had its size changed. This caused exceptions when flushing the texture. Now it correctly takes the block size into account, assuming that these textures are only related because a pixel in the non-compressed texture represents a block in the compressed one. * Implement JD's suggestion for HashCode Combine Co-authored-by: jduncanator <1518948+jduncanator@users.noreply.github.com> * Address feedback * Address feedback. Co-authored-by: jduncanator <1518948+jduncanator@users.noreply.github.com>
1160 lines
No EOL
44 KiB
C#
1160 lines
No EOL
44 KiB
C#
using Ryujinx.Common.Logging;
|
|
using Ryujinx.Graphics.GAL;
|
|
using Ryujinx.Graphics.Gpu.Image;
|
|
using Ryujinx.Graphics.Gpu.Memory;
|
|
using Ryujinx.Graphics.Gpu.Shader;
|
|
using Ryujinx.Graphics.Gpu.State;
|
|
using Ryujinx.Graphics.Shader;
|
|
using System;
|
|
using System.Runtime.InteropServices;
|
|
|
|
namespace Ryujinx.Graphics.Gpu.Engine
|
|
{
|
|
using Texture = Image.Texture;
|
|
|
|
/// <summary>
|
|
/// GPU method implementations.
|
|
/// </summary>
|
|
partial class Methods
|
|
{
|
|
private readonly GpuContext _context;
|
|
private readonly ShaderProgramInfo[] _currentProgramInfo;
|
|
|
|
/// <summary>
|
|
/// In-memory shader cache.
|
|
/// </summary>
|
|
public ShaderCache ShaderCache { get; }
|
|
|
|
/// <summary>
|
|
/// GPU buffer manager.
|
|
/// </summary>
|
|
public BufferManager BufferManager { get; }
|
|
|
|
/// <summary>
|
|
/// GPU texture manager.
|
|
/// </summary>
|
|
public TextureManager TextureManager { get; }
|
|
|
|
private bool _isAnyVbInstanced;
|
|
private bool _vsUsesInstanceId;
|
|
|
|
private bool _forceShaderUpdate;
|
|
|
|
private bool _prevTfEnable;
|
|
|
|
/// <summary>
|
|
/// Creates a new instance of the GPU methods class.
|
|
/// </summary>
|
|
/// <param name="context">GPU context</param>
|
|
public Methods(GpuContext context)
|
|
{
|
|
_context = context;
|
|
|
|
ShaderCache = new ShaderCache(_context);
|
|
|
|
_currentProgramInfo = new ShaderProgramInfo[Constants.ShaderStages];
|
|
|
|
BufferManager = new BufferManager(context);
|
|
TextureManager = new TextureManager(context);
|
|
|
|
context.MemoryManager.MemoryUnmapped += _counterCache.MemoryUnmappedHandler;
|
|
context.MemoryManager.MemoryUnmapped += TextureManager.MemoryUnmappedHandler;
|
|
}
|
|
|
|
/// <summary>
|
|
/// Register callback for GPU method calls that triggers an action on the GPU.
|
|
/// </summary>
|
|
/// <param name="state">GPU state where the triggers will be registered</param>
|
|
public void RegisterCallbacks(GpuState state)
|
|
{
|
|
state.RegisterCallback(MethodOffset.LaunchDma, LaunchDma);
|
|
state.RegisterCallback(MethodOffset.LoadInlineData, LoadInlineData);
|
|
|
|
state.RegisterCallback(MethodOffset.Dispatch, Dispatch);
|
|
|
|
state.RegisterCallback(MethodOffset.SyncpointAction, IncrementSyncpoint);
|
|
|
|
state.RegisterCallback(MethodOffset.CopyBuffer, CopyBuffer);
|
|
state.RegisterCallback(MethodOffset.CopyTexture, CopyTexture);
|
|
|
|
state.RegisterCallback(MethodOffset.TextureBarrier, TextureBarrier);
|
|
state.RegisterCallback(MethodOffset.InvalidateTextures, InvalidateTextures);
|
|
state.RegisterCallback(MethodOffset.TextureBarrierTiled, TextureBarrierTiled);
|
|
|
|
state.RegisterCallback(MethodOffset.VbElementU8, VbElementU8);
|
|
state.RegisterCallback(MethodOffset.VbElementU16, VbElementU16);
|
|
state.RegisterCallback(MethodOffset.VbElementU32, VbElementU32);
|
|
|
|
state.RegisterCallback(MethodOffset.ResetCounter, ResetCounter);
|
|
|
|
state.RegisterCallback(MethodOffset.DrawEnd, DrawEnd);
|
|
state.RegisterCallback(MethodOffset.DrawBegin, DrawBegin);
|
|
|
|
state.RegisterCallback(MethodOffset.IndexBufferCount, SetIndexBufferCount);
|
|
|
|
state.RegisterCallback(MethodOffset.Clear, Clear);
|
|
|
|
state.RegisterCallback(MethodOffset.Report, Report);
|
|
|
|
state.RegisterCallback(MethodOffset.FirmwareCall4, FirmwareCall4);
|
|
|
|
state.RegisterCallback(MethodOffset.UniformBufferUpdateData, 16, UniformBufferUpdate);
|
|
|
|
state.RegisterCallback(MethodOffset.UniformBufferBindVertex, UniformBufferBindVertex);
|
|
state.RegisterCallback(MethodOffset.UniformBufferBindTessControl, UniformBufferBindTessControl);
|
|
state.RegisterCallback(MethodOffset.UniformBufferBindTessEvaluation, UniformBufferBindTessEvaluation);
|
|
state.RegisterCallback(MethodOffset.UniformBufferBindGeometry, UniformBufferBindGeometry);
|
|
state.RegisterCallback(MethodOffset.UniformBufferBindFragment, UniformBufferBindFragment);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host state based on the current guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Guest GPU state</param>
|
|
private void UpdateState(GpuState state)
|
|
{
|
|
bool tfEnable = state.Get<Boolean32>(MethodOffset.TfEnable);
|
|
|
|
if (!tfEnable && _prevTfEnable)
|
|
{
|
|
_context.Renderer.Pipeline.EndTransformFeedback();
|
|
_prevTfEnable = false;
|
|
}
|
|
|
|
// Shaders must be the first one to be updated if modified, because
|
|
// some of the other state depends on information from the currently
|
|
// bound shaders.
|
|
if (state.QueryModified(MethodOffset.ShaderBaseAddress, MethodOffset.ShaderState) || _forceShaderUpdate)
|
|
{
|
|
_forceShaderUpdate = false;
|
|
|
|
UpdateShaderState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.TfBufferState))
|
|
{
|
|
UpdateTfBufferState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.ClipDistanceEnable))
|
|
{
|
|
UpdateUserClipState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.RasterizeEnable))
|
|
{
|
|
UpdateRasterizerState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.RtColorState,
|
|
MethodOffset.RtDepthStencilState,
|
|
MethodOffset.RtControl,
|
|
MethodOffset.RtDepthStencilSize,
|
|
MethodOffset.RtDepthStencilEnable))
|
|
{
|
|
UpdateRenderTargetState(state, useControl: true);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.ScissorState))
|
|
{
|
|
UpdateScissorState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.ViewVolumeClipControl))
|
|
{
|
|
UpdateDepthClampState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.AlphaTestEnable,
|
|
MethodOffset.AlphaTestRef,
|
|
MethodOffset.AlphaTestFunc))
|
|
{
|
|
UpdateAlphaTestState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.DepthTestEnable,
|
|
MethodOffset.DepthWriteEnable,
|
|
MethodOffset.DepthTestFunc))
|
|
{
|
|
UpdateDepthTestState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.DepthMode,
|
|
MethodOffset.ViewportTransform,
|
|
MethodOffset.ViewportExtents))
|
|
{
|
|
UpdateViewportTransform(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.DepthBiasState,
|
|
MethodOffset.DepthBiasFactor,
|
|
MethodOffset.DepthBiasUnits,
|
|
MethodOffset.DepthBiasClamp))
|
|
{
|
|
UpdateDepthBiasState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.StencilBackMasks,
|
|
MethodOffset.StencilTestState,
|
|
MethodOffset.StencilBackTestState))
|
|
{
|
|
UpdateStencilTestState(state);
|
|
}
|
|
|
|
// Pools.
|
|
if (state.QueryModified(MethodOffset.SamplerPoolState, MethodOffset.SamplerIndex))
|
|
{
|
|
UpdateSamplerPoolState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.TexturePoolState))
|
|
{
|
|
UpdateTexturePoolState(state);
|
|
}
|
|
|
|
// Input assembler state.
|
|
if (state.QueryModified(MethodOffset.VertexAttribState))
|
|
{
|
|
UpdateVertexAttribState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.PointSize,
|
|
MethodOffset.VertexProgramPointSize,
|
|
MethodOffset.PointSpriteEnable,
|
|
MethodOffset.PointCoordReplace))
|
|
{
|
|
UpdatePointState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.PrimitiveRestartState))
|
|
{
|
|
UpdatePrimitiveRestartState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.IndexBufferState))
|
|
{
|
|
UpdateIndexBufferState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.VertexBufferDrawState,
|
|
MethodOffset.VertexBufferInstanced,
|
|
MethodOffset.VertexBufferState,
|
|
MethodOffset.VertexBufferEndAddress))
|
|
{
|
|
UpdateVertexBufferState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.FaceState))
|
|
{
|
|
UpdateFaceState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.RtColorMaskShared, MethodOffset.RtColorMask))
|
|
{
|
|
UpdateRtColorMask(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.BlendIndependent,
|
|
MethodOffset.BlendConstant,
|
|
MethodOffset.BlendStateCommon,
|
|
MethodOffset.BlendEnableCommon,
|
|
MethodOffset.BlendEnable,
|
|
MethodOffset.BlendState))
|
|
{
|
|
UpdateBlendState(state);
|
|
}
|
|
|
|
if (state.QueryModified(MethodOffset.LogicOpState))
|
|
{
|
|
UpdateLogicOpState(state);
|
|
}
|
|
|
|
CommitBindings();
|
|
|
|
if (tfEnable && !_prevTfEnable)
|
|
{
|
|
_context.Renderer.Pipeline.BeginTransformFeedback(PrimitiveType.Convert());
|
|
_prevTfEnable = true;
|
|
}
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates Rasterizer primitive discard state based on guest gpu state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateRasterizerState(GpuState state)
|
|
{
|
|
Boolean32 enable = state.Get<Boolean32>(MethodOffset.RasterizeEnable);
|
|
_context.Renderer.Pipeline.SetRasterizerDiscard(!enable);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Ensures that the bindings are visible to the host GPU.
|
|
/// Note: this actually performs the binding using the host graphics API.
|
|
/// </summary>
|
|
private void CommitBindings()
|
|
{
|
|
UpdateStorageBuffers();
|
|
|
|
BufferManager.CommitGraphicsBindings();
|
|
TextureManager.CommitGraphicsBindings();
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates storage buffer bindings.
|
|
/// </summary>
|
|
private void UpdateStorageBuffers()
|
|
{
|
|
for (int stage = 0; stage < _currentProgramInfo.Length; stage++)
|
|
{
|
|
ShaderProgramInfo info = _currentProgramInfo[stage];
|
|
|
|
if (info == null)
|
|
{
|
|
continue;
|
|
}
|
|
|
|
for (int index = 0; index < info.SBuffers.Count; index++)
|
|
{
|
|
BufferDescriptor sb = info.SBuffers[index];
|
|
|
|
ulong sbDescAddress = BufferManager.GetGraphicsUniformBufferAddress(stage, 0);
|
|
|
|
int sbDescOffset = 0x110 + stage * 0x100 + sb.Slot * 0x10;
|
|
|
|
sbDescAddress += (ulong)sbDescOffset;
|
|
|
|
SbDescriptor sbDescriptor = _context.PhysicalMemory.Read<SbDescriptor>(sbDescAddress);
|
|
|
|
BufferManager.SetGraphicsStorageBuffer(stage, sb.Slot, sbDescriptor.PackAddress(), (uint)sbDescriptor.Size);
|
|
}
|
|
}
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates render targets (color and depth-stencil buffers) based on current render target state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
/// <param name="useControl">Use draw buffers information from render target control register</param>
|
|
/// <param name="singleUse">If this is not -1, it indicates that only the given indexed target will be used.</param>
|
|
private void UpdateRenderTargetState(GpuState state, bool useControl, int singleUse = -1)
|
|
{
|
|
var rtControl = state.Get<RtControl>(MethodOffset.RtControl);
|
|
|
|
int count = useControl ? rtControl.UnpackCount() : Constants.TotalRenderTargets;
|
|
|
|
var msaaMode = state.Get<TextureMsaaMode>(MethodOffset.RtMsaaMode);
|
|
|
|
int samplesInX = msaaMode.SamplesInX();
|
|
int samplesInY = msaaMode.SamplesInY();
|
|
|
|
bool changedScale = false;
|
|
|
|
for (int index = 0; index < Constants.TotalRenderTargets; index++)
|
|
{
|
|
int rtIndex = useControl ? rtControl.UnpackPermutationIndex(index) : index;
|
|
|
|
var colorState = state.Get<RtColorState>(MethodOffset.RtColorState, rtIndex);
|
|
|
|
if (index >= count || !IsRtEnabled(colorState))
|
|
{
|
|
changedScale |= TextureManager.SetRenderTargetColor(index, null);
|
|
|
|
continue;
|
|
}
|
|
|
|
Texture color = TextureManager.FindOrCreateTexture(colorState, samplesInX, samplesInY);
|
|
|
|
changedScale |= TextureManager.SetRenderTargetColor(index, color);
|
|
|
|
if (color != null)
|
|
{
|
|
color.SignalModified();
|
|
}
|
|
}
|
|
|
|
bool dsEnable = state.Get<Boolean32>(MethodOffset.RtDepthStencilEnable);
|
|
|
|
Texture depthStencil = null;
|
|
|
|
if (dsEnable)
|
|
{
|
|
var dsState = state.Get<RtDepthStencilState>(MethodOffset.RtDepthStencilState);
|
|
var dsSize = state.Get<Size3D>(MethodOffset.RtDepthStencilSize);
|
|
|
|
depthStencil = TextureManager.FindOrCreateTexture(dsState, dsSize, samplesInX, samplesInY);
|
|
}
|
|
|
|
changedScale |= TextureManager.SetRenderTargetDepthStencil(depthStencil);
|
|
|
|
if (changedScale)
|
|
{
|
|
TextureManager.UpdateRenderTargetScale(singleUse);
|
|
_context.Renderer.Pipeline.SetRenderTargetScale(TextureManager.RenderTargetScale);
|
|
|
|
UpdateViewportTransform(state);
|
|
UpdateScissorState(state);
|
|
}
|
|
|
|
if (depthStencil != null)
|
|
{
|
|
depthStencil.SignalModified();
|
|
}
|
|
}
|
|
|
|
/// <summary>
|
|
/// Checks if a render target color buffer is used.
|
|
/// </summary>
|
|
/// <param name="colorState">Color buffer information</param>
|
|
/// <returns>True if the specified buffer is enabled/used, false otherwise</returns>
|
|
private static bool IsRtEnabled(RtColorState colorState)
|
|
{
|
|
// Colors are disabled by writing 0 to the format.
|
|
return colorState.Format != 0 && colorState.WidthOrStride != 0;
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host scissor test state based on current GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateScissorState(GpuState state)
|
|
{
|
|
for (int index = 0; index < Constants.TotalViewports; index++)
|
|
{
|
|
ScissorState scissor = state.Get<ScissorState>(MethodOffset.ScissorState, index);
|
|
|
|
bool enable = scissor.Enable && (scissor.X1 != 0 || scissor.Y1 != 0 || scissor.X2 != 0xffff || scissor.Y2 != 0xffff);
|
|
|
|
_context.Renderer.Pipeline.SetScissorEnable(index, enable);
|
|
|
|
if (enable)
|
|
{
|
|
int x = scissor.X1;
|
|
int y = scissor.Y1;
|
|
int width = scissor.X2 - x;
|
|
int height = scissor.Y2 - y;
|
|
|
|
float scale = TextureManager.RenderTargetScale;
|
|
if (scale != 1f)
|
|
{
|
|
x = (int)(x * scale);
|
|
y = (int)(y * scale);
|
|
width = (int)Math.Ceiling(width * scale);
|
|
height = (int)Math.Ceiling(height * scale);
|
|
}
|
|
|
|
_context.Renderer.Pipeline.SetScissor(index, x, y, width, height);
|
|
}
|
|
}
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host depth clamp state based on current GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateDepthClampState(GpuState state)
|
|
{
|
|
ViewVolumeClipControl clip = state.Get<ViewVolumeClipControl>(MethodOffset.ViewVolumeClipControl);
|
|
_context.Renderer.Pipeline.SetDepthClamp((clip & ViewVolumeClipControl.DepthClampDisabled) == 0);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host alpha test state based on current GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateAlphaTestState(GpuState state)
|
|
{
|
|
_context.Renderer.Pipeline.SetAlphaTest(
|
|
state.Get<Boolean32>(MethodOffset.AlphaTestEnable),
|
|
state.Get<float>(MethodOffset.AlphaTestRef),
|
|
state.Get<CompareOp>(MethodOffset.AlphaTestFunc));
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host depth test state based on current GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateDepthTestState(GpuState state)
|
|
{
|
|
_context.Renderer.Pipeline.SetDepthTest(new DepthTestDescriptor(
|
|
state.Get<Boolean32>(MethodOffset.DepthTestEnable),
|
|
state.Get<Boolean32>(MethodOffset.DepthWriteEnable),
|
|
state.Get<CompareOp>(MethodOffset.DepthTestFunc)));
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host viewport transform and clipping state based on current GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateViewportTransform(GpuState state)
|
|
{
|
|
DepthMode depthMode = state.Get<DepthMode>(MethodOffset.DepthMode);
|
|
|
|
_context.Renderer.Pipeline.SetDepthMode(depthMode);
|
|
|
|
YControl yControl = state.Get<YControl>(MethodOffset.YControl);
|
|
|
|
bool flipY = yControl.HasFlag(YControl.NegateY);
|
|
Origin origin = yControl.HasFlag(YControl.TriangleRastFlip) ? Origin.LowerLeft : Origin.UpperLeft;
|
|
|
|
_context.Renderer.Pipeline.SetOrigin(origin);
|
|
|
|
// The triangle rast flip flag only affects rasterization, the viewport is not flipped.
|
|
// Setting the origin mode to upper left on the host, however, not only affects rasterization,
|
|
// but also flips the viewport.
|
|
// We negate the effects of flipping the viewport by flipping it again using the viewport swizzle.
|
|
if (origin == Origin.UpperLeft)
|
|
{
|
|
flipY = !flipY;
|
|
}
|
|
|
|
Span<Viewport> viewports = stackalloc Viewport[Constants.TotalViewports];
|
|
|
|
for (int index = 0; index < Constants.TotalViewports; index++)
|
|
{
|
|
var transform = state.Get<ViewportTransform>(MethodOffset.ViewportTransform, index);
|
|
var extents = state.Get<ViewportExtents> (MethodOffset.ViewportExtents, index);
|
|
|
|
float x = transform.TranslateX - MathF.Abs(transform.ScaleX);
|
|
float y = transform.TranslateY - MathF.Abs(transform.ScaleY);
|
|
|
|
float width = MathF.Abs(transform.ScaleX) * 2;
|
|
float height = MathF.Abs(transform.ScaleY) * 2;
|
|
|
|
float scale = TextureManager.RenderTargetScale;
|
|
if (scale != 1f)
|
|
{
|
|
x *= scale;
|
|
y *= scale;
|
|
width *= scale;
|
|
height *= scale;
|
|
}
|
|
|
|
RectangleF region = new RectangleF(x, y, width, height);
|
|
|
|
ViewportSwizzle swizzleX = transform.UnpackSwizzleX();
|
|
ViewportSwizzle swizzleY = transform.UnpackSwizzleY();
|
|
ViewportSwizzle swizzleZ = transform.UnpackSwizzleZ();
|
|
ViewportSwizzle swizzleW = transform.UnpackSwizzleW();
|
|
|
|
if (transform.ScaleX < 0)
|
|
{
|
|
swizzleX ^= ViewportSwizzle.NegativeFlag;
|
|
}
|
|
|
|
if (flipY)
|
|
{
|
|
swizzleY ^= ViewportSwizzle.NegativeFlag;
|
|
}
|
|
|
|
if (transform.ScaleY < 0)
|
|
{
|
|
swizzleY ^= ViewportSwizzle.NegativeFlag;
|
|
}
|
|
|
|
if (transform.ScaleZ < 0)
|
|
{
|
|
swizzleZ ^= ViewportSwizzle.NegativeFlag;
|
|
}
|
|
|
|
viewports[index] = new Viewport(
|
|
region,
|
|
swizzleX,
|
|
swizzleY,
|
|
swizzleZ,
|
|
swizzleW,
|
|
extents.DepthNear,
|
|
extents.DepthFar);
|
|
}
|
|
|
|
_context.Renderer.Pipeline.SetViewports(0, viewports);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host depth bias (also called polygon offset) state based on current GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateDepthBiasState(GpuState state)
|
|
{
|
|
var depthBias = state.Get<DepthBiasState>(MethodOffset.DepthBiasState);
|
|
|
|
float factor = state.Get<float>(MethodOffset.DepthBiasFactor);
|
|
float units = state.Get<float>(MethodOffset.DepthBiasUnits);
|
|
float clamp = state.Get<float>(MethodOffset.DepthBiasClamp);
|
|
|
|
PolygonModeMask enables;
|
|
|
|
enables = (depthBias.PointEnable ? PolygonModeMask.Point : 0);
|
|
enables |= (depthBias.LineEnable ? PolygonModeMask.Line : 0);
|
|
enables |= (depthBias.FillEnable ? PolygonModeMask.Fill : 0);
|
|
|
|
_context.Renderer.Pipeline.SetDepthBias(enables, factor, units / 2f, clamp);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host stencil test state based on current GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateStencilTestState(GpuState state)
|
|
{
|
|
var backMasks = state.Get<StencilBackMasks>(MethodOffset.StencilBackMasks);
|
|
var test = state.Get<StencilTestState>(MethodOffset.StencilTestState);
|
|
var backTest = state.Get<StencilBackTestState>(MethodOffset.StencilBackTestState);
|
|
|
|
CompareOp backFunc;
|
|
StencilOp backSFail;
|
|
StencilOp backDpPass;
|
|
StencilOp backDpFail;
|
|
int backFuncRef;
|
|
int backFuncMask;
|
|
int backMask;
|
|
|
|
if (backTest.TwoSided)
|
|
{
|
|
backFunc = backTest.BackFunc;
|
|
backSFail = backTest.BackSFail;
|
|
backDpPass = backTest.BackDpPass;
|
|
backDpFail = backTest.BackDpFail;
|
|
backFuncRef = backMasks.FuncRef;
|
|
backFuncMask = backMasks.FuncMask;
|
|
backMask = backMasks.Mask;
|
|
}
|
|
else
|
|
{
|
|
backFunc = test.FrontFunc;
|
|
backSFail = test.FrontSFail;
|
|
backDpPass = test.FrontDpPass;
|
|
backDpFail = test.FrontDpFail;
|
|
backFuncRef = test.FrontFuncRef;
|
|
backFuncMask = test.FrontFuncMask;
|
|
backMask = test.FrontMask;
|
|
}
|
|
|
|
_context.Renderer.Pipeline.SetStencilTest(new StencilTestDescriptor(
|
|
test.Enable,
|
|
test.FrontFunc,
|
|
test.FrontSFail,
|
|
test.FrontDpPass,
|
|
test.FrontDpFail,
|
|
test.FrontFuncRef,
|
|
test.FrontFuncMask,
|
|
test.FrontMask,
|
|
backFunc,
|
|
backSFail,
|
|
backDpPass,
|
|
backDpFail,
|
|
backFuncRef,
|
|
backFuncMask,
|
|
backMask));
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates current sampler pool address and size based on guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateSamplerPoolState(GpuState state)
|
|
{
|
|
var texturePool = state.Get<PoolState>(MethodOffset.TexturePoolState);
|
|
var samplerPool = state.Get<PoolState>(MethodOffset.SamplerPoolState);
|
|
|
|
var samplerIndex = state.Get<SamplerIndex>(MethodOffset.SamplerIndex);
|
|
|
|
int maximumId = samplerIndex == SamplerIndex.ViaHeaderIndex
|
|
? texturePool.MaximumId
|
|
: samplerPool.MaximumId;
|
|
|
|
TextureManager.SetGraphicsSamplerPool(samplerPool.Address.Pack(), maximumId, samplerIndex);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates current texture pool address and size based on guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateTexturePoolState(GpuState state)
|
|
{
|
|
var texturePool = state.Get<PoolState>(MethodOffset.TexturePoolState);
|
|
|
|
TextureManager.SetGraphicsTexturePool(texturePool.Address.Pack(), texturePool.MaximumId);
|
|
|
|
TextureManager.SetGraphicsTextureBufferIndex(state.Get<int>(MethodOffset.TextureBufferIndex));
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host vertex attributes based on guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateVertexAttribState(GpuState state)
|
|
{
|
|
Span<VertexAttribDescriptor> vertexAttribs = stackalloc VertexAttribDescriptor[Constants.TotalVertexAttribs];
|
|
|
|
for (int index = 0; index < Constants.TotalVertexAttribs; index++)
|
|
{
|
|
var vertexAttrib = state.Get<VertexAttribState>(MethodOffset.VertexAttribState, index);
|
|
|
|
if (!FormatTable.TryGetAttribFormat(vertexAttrib.UnpackFormat(), out Format format))
|
|
{
|
|
Logger.Debug?.Print(LogClass.Gpu, $"Invalid attribute format 0x{vertexAttrib.UnpackFormat():X}.");
|
|
|
|
format = Format.R32G32B32A32Float;
|
|
}
|
|
|
|
vertexAttribs[index] = new VertexAttribDescriptor(
|
|
vertexAttrib.UnpackBufferIndex(),
|
|
vertexAttrib.UnpackOffset(),
|
|
vertexAttrib.UnpackIsConstant(),
|
|
format);
|
|
}
|
|
|
|
_context.Renderer.Pipeline.SetVertexAttribs(vertexAttribs);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host point size based on guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdatePointState(GpuState state)
|
|
{
|
|
float size = state.Get<float>(MethodOffset.PointSize);
|
|
bool isProgramPointSize = state.Get<Boolean32>(MethodOffset.VertexProgramPointSize);
|
|
bool enablePointSprite = state.Get<Boolean32>(MethodOffset.PointSpriteEnable);
|
|
|
|
// TODO: Need to figure out a way to map PointCoordReplace enable bit.
|
|
Origin origin = (state.Get<int>(MethodOffset.PointCoordReplace) & 4) == 0 ? Origin.LowerLeft : Origin.UpperLeft;
|
|
|
|
_context.Renderer.Pipeline.SetPointParameters(size, isProgramPointSize, enablePointSprite, origin);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host primitive restart based on guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdatePrimitiveRestartState(GpuState state)
|
|
{
|
|
PrimitiveRestartState primitiveRestart = state.Get<PrimitiveRestartState>(MethodOffset.PrimitiveRestartState);
|
|
|
|
_context.Renderer.Pipeline.SetPrimitiveRestart(
|
|
primitiveRestart.Enable,
|
|
primitiveRestart.Index);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host index buffer binding based on guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateIndexBufferState(GpuState state)
|
|
{
|
|
var indexBuffer = state.Get<IndexBufferState>(MethodOffset.IndexBufferState);
|
|
|
|
_firstIndex = indexBuffer.First;
|
|
_indexCount = indexBuffer.Count;
|
|
|
|
if (_indexCount == 0)
|
|
{
|
|
return;
|
|
}
|
|
|
|
ulong gpuVa = indexBuffer.Address.Pack();
|
|
|
|
// Do not use the end address to calculate the size, because
|
|
// the result may be much larger than the real size of the index buffer.
|
|
ulong size = (ulong)(_firstIndex + _indexCount);
|
|
|
|
switch (indexBuffer.Type)
|
|
{
|
|
case IndexType.UShort: size *= 2; break;
|
|
case IndexType.UInt: size *= 4; break;
|
|
}
|
|
|
|
BufferManager.SetIndexBuffer(gpuVa, size, indexBuffer.Type);
|
|
|
|
// The index buffer affects the vertex buffer size calculation, we
|
|
// need to ensure that they are updated.
|
|
UpdateVertexBufferState(state);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host vertex buffer bindings based on guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateVertexBufferState(GpuState state)
|
|
{
|
|
_isAnyVbInstanced = false;
|
|
|
|
for (int index = 0; index < Constants.TotalVertexBuffers; index++)
|
|
{
|
|
var vertexBuffer = state.Get<VertexBufferState>(MethodOffset.VertexBufferState, index);
|
|
|
|
if (!vertexBuffer.UnpackEnable())
|
|
{
|
|
BufferManager.SetVertexBuffer(index, 0, 0, 0, 0);
|
|
|
|
continue;
|
|
}
|
|
|
|
GpuVa endAddress = state.Get<GpuVa>(MethodOffset.VertexBufferEndAddress, index);
|
|
|
|
ulong address = vertexBuffer.Address.Pack();
|
|
|
|
int stride = vertexBuffer.UnpackStride();
|
|
|
|
bool instanced = state.Get<Boolean32>(MethodOffset.VertexBufferInstanced + index);
|
|
|
|
int divisor = instanced ? vertexBuffer.Divisor : 0;
|
|
|
|
_isAnyVbInstanced |= divisor != 0;
|
|
|
|
ulong size;
|
|
|
|
if (_inlineIndexCount != 0 || _drawIndexed || stride == 0 || instanced)
|
|
{
|
|
// This size may be (much) larger than the real vertex buffer size.
|
|
// Avoid calculating it this way, unless we don't have any other option.
|
|
size = endAddress.Pack() - address + 1;
|
|
}
|
|
else
|
|
{
|
|
// For non-indexed draws, we can guess the size from the vertex count
|
|
// and stride.
|
|
int firstInstance = state.Get<int>(MethodOffset.FirstInstance);
|
|
|
|
var drawState = state.Get<VertexBufferDrawState>(MethodOffset.VertexBufferDrawState);
|
|
|
|
size = (ulong)((firstInstance + drawState.First + drawState.Count) * stride);
|
|
}
|
|
|
|
BufferManager.SetVertexBuffer(index, address, size, stride, divisor);
|
|
}
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host face culling and orientation based on guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateFaceState(GpuState state)
|
|
{
|
|
var face = state.Get<FaceState>(MethodOffset.FaceState);
|
|
|
|
_context.Renderer.Pipeline.SetFaceCulling(face.CullEnable, face.CullFace);
|
|
|
|
_context.Renderer.Pipeline.SetFrontFace(face.FrontFace);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host render target color masks, based on guest GPU state.
|
|
/// This defines which color channels are written to each color buffer.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateRtColorMask(GpuState state)
|
|
{
|
|
bool rtColorMaskShared = state.Get<Boolean32>(MethodOffset.RtColorMaskShared);
|
|
|
|
Span<uint> componentMasks = stackalloc uint[Constants.TotalRenderTargets];
|
|
|
|
for (int index = 0; index < Constants.TotalRenderTargets; index++)
|
|
{
|
|
var colorMask = state.Get<RtColorMask>(MethodOffset.RtColorMask, rtColorMaskShared ? 0 : index);
|
|
|
|
uint componentMask;
|
|
|
|
componentMask = (colorMask.UnpackRed() ? 1u : 0u);
|
|
componentMask |= (colorMask.UnpackGreen() ? 2u : 0u);
|
|
componentMask |= (colorMask.UnpackBlue() ? 4u : 0u);
|
|
componentMask |= (colorMask.UnpackAlpha() ? 8u : 0u);
|
|
|
|
componentMasks[index] = componentMask;
|
|
}
|
|
|
|
_context.Renderer.Pipeline.SetRenderTargetColorMasks(componentMasks);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host render target color buffer blending state, based on guest state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateBlendState(GpuState state)
|
|
{
|
|
bool blendIndependent = state.Get<Boolean32>(MethodOffset.BlendIndependent);
|
|
ColorF blendConstant = state.Get<ColorF>(MethodOffset.BlendConstant);
|
|
|
|
for (int index = 0; index < Constants.TotalRenderTargets; index++)
|
|
{
|
|
BlendDescriptor descriptor;
|
|
|
|
if (blendIndependent)
|
|
{
|
|
bool enable = state.Get<Boolean32> (MethodOffset.BlendEnable, index);
|
|
var blend = state.Get<BlendState>(MethodOffset.BlendState, index);
|
|
|
|
descriptor = new BlendDescriptor(
|
|
enable,
|
|
blendConstant,
|
|
blend.ColorOp,
|
|
blend.ColorSrcFactor,
|
|
blend.ColorDstFactor,
|
|
blend.AlphaOp,
|
|
blend.AlphaSrcFactor,
|
|
blend.AlphaDstFactor);
|
|
}
|
|
else
|
|
{
|
|
bool enable = state.Get<Boolean32> (MethodOffset.BlendEnable, 0);
|
|
var blend = state.Get<BlendStateCommon>(MethodOffset.BlendStateCommon);
|
|
|
|
descriptor = new BlendDescriptor(
|
|
enable,
|
|
blendConstant,
|
|
blend.ColorOp,
|
|
blend.ColorSrcFactor,
|
|
blend.ColorDstFactor,
|
|
blend.AlphaOp,
|
|
blend.AlphaSrcFactor,
|
|
blend.AlphaDstFactor);
|
|
}
|
|
|
|
_context.Renderer.Pipeline.SetBlendState(index, descriptor);
|
|
}
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host logical operation state, based on guest state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
public void UpdateLogicOpState(GpuState state)
|
|
{
|
|
LogicalOpState logicOpState = state.Get<LogicalOpState>(MethodOffset.LogicOpState);
|
|
|
|
_context.Renderer.Pipeline.SetLogicOpState(logicOpState.Enable, logicOpState.LogicalOp);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Storage buffer address and size information.
|
|
/// </summary>
|
|
private struct SbDescriptor
|
|
{
|
|
#pragma warning disable CS0649
|
|
public uint AddressLow;
|
|
public uint AddressHigh;
|
|
public int Size;
|
|
public int Padding;
|
|
#pragma warning restore CS0649
|
|
|
|
public ulong PackAddress()
|
|
{
|
|
return AddressLow | ((ulong)AddressHigh << 32);
|
|
}
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates host shaders based on the guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateShaderState(GpuState state)
|
|
{
|
|
ShaderAddresses addresses = new ShaderAddresses();
|
|
|
|
Span<ShaderAddresses> addressesSpan = MemoryMarshal.CreateSpan(ref addresses, 1);
|
|
|
|
Span<ulong> addressesArray = MemoryMarshal.Cast<ShaderAddresses, ulong>(addressesSpan);
|
|
|
|
ulong baseAddress = state.Get<GpuVa>(MethodOffset.ShaderBaseAddress).Pack();
|
|
|
|
for (int index = 0; index < 6; index++)
|
|
{
|
|
var shader = state.Get<ShaderState>(MethodOffset.ShaderState, index);
|
|
|
|
if (!shader.UnpackEnable() && index != 1)
|
|
{
|
|
continue;
|
|
}
|
|
|
|
addressesArray[index] = baseAddress + shader.Offset;
|
|
}
|
|
|
|
ShaderBundle gs = ShaderCache.GetGraphicsShader(state, addresses);
|
|
|
|
_vsUsesInstanceId = gs.Shaders[0]?.Program.Info.UsesInstanceId ?? false;
|
|
|
|
for (int stage = 0; stage < Constants.ShaderStages; stage++)
|
|
{
|
|
ShaderProgramInfo info = gs.Shaders[stage]?.Program.Info;
|
|
|
|
_currentProgramInfo[stage] = info;
|
|
|
|
if (info == null)
|
|
{
|
|
continue;
|
|
}
|
|
|
|
var textureBindings = new TextureBindingInfo[info.Textures.Count];
|
|
|
|
for (int index = 0; index < info.Textures.Count; index++)
|
|
{
|
|
var descriptor = info.Textures[index];
|
|
|
|
Target target = GetTarget(descriptor.Type);
|
|
|
|
if (descriptor.IsBindless)
|
|
{
|
|
textureBindings[index] = new TextureBindingInfo(target, descriptor.CbufSlot, descriptor.CbufOffset, descriptor.Flags);
|
|
}
|
|
else
|
|
{
|
|
textureBindings[index] = new TextureBindingInfo(target, descriptor.HandleIndex, descriptor.Flags);
|
|
}
|
|
}
|
|
|
|
TextureManager.SetGraphicsTextures(stage, textureBindings);
|
|
|
|
var imageBindings = new TextureBindingInfo[info.Images.Count];
|
|
|
|
for (int index = 0; index < info.Images.Count; index++)
|
|
{
|
|
var descriptor = info.Images[index];
|
|
|
|
Target target = GetTarget(descriptor.Type);
|
|
|
|
imageBindings[index] = new TextureBindingInfo(target, descriptor.HandleIndex, descriptor.Flags);
|
|
}
|
|
|
|
TextureManager.SetGraphicsImages(stage, imageBindings);
|
|
|
|
uint sbEnableMask = 0;
|
|
uint ubEnableMask = 0;
|
|
|
|
for (int index = 0; index < info.SBuffers.Count; index++)
|
|
{
|
|
sbEnableMask |= 1u << info.SBuffers[index].Slot;
|
|
}
|
|
|
|
for (int index = 0; index < info.CBuffers.Count; index++)
|
|
{
|
|
ubEnableMask |= 1u << info.CBuffers[index].Slot;
|
|
}
|
|
|
|
BufferManager.SetGraphicsStorageBufferEnableMask(stage, sbEnableMask);
|
|
BufferManager.SetGraphicsUniformBufferEnableMask(stage, ubEnableMask);
|
|
}
|
|
|
|
_context.Renderer.Pipeline.SetProgram(gs.HostProgram);
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates transform feedback buffer state based on the guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateTfBufferState(GpuState state)
|
|
{
|
|
for (int index = 0; index < Constants.TotalTransformFeedbackBuffers; index++)
|
|
{
|
|
TfBufferState tfb = state.Get<TfBufferState>(MethodOffset.TfBufferState, index);
|
|
|
|
if (!tfb.Enable)
|
|
{
|
|
BufferManager.SetTransformFeedbackBuffer(index, 0, 0);
|
|
|
|
continue;
|
|
}
|
|
|
|
BufferManager.SetTransformFeedbackBuffer(index, tfb.Address.Pack(), (uint)tfb.Size);
|
|
}
|
|
}
|
|
|
|
/// <summary>
|
|
/// Updates user-defined clipping based on the guest GPU state.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state</param>
|
|
private void UpdateUserClipState(GpuState state)
|
|
{
|
|
int clipMask = state.Get<int>(MethodOffset.ClipDistanceEnable);
|
|
|
|
for (int i = 0; i < Constants.TotalClipDistances; ++i)
|
|
{
|
|
_context.Renderer.Pipeline.SetUserClipDistance(i, (clipMask & (1 << i)) != 0);
|
|
}
|
|
}
|
|
|
|
/// <summary>
|
|
/// Gets texture target from a sampler type.
|
|
/// </summary>
|
|
/// <param name="type">Sampler type</param>
|
|
/// <returns>Texture target value</returns>
|
|
private static Target GetTarget(SamplerType type)
|
|
{
|
|
type &= ~(SamplerType.Indexed | SamplerType.Shadow);
|
|
|
|
switch (type)
|
|
{
|
|
case SamplerType.Texture1D:
|
|
return Target.Texture1D;
|
|
|
|
case SamplerType.TextureBuffer:
|
|
return Target.TextureBuffer;
|
|
|
|
case SamplerType.Texture1D | SamplerType.Array:
|
|
return Target.Texture1DArray;
|
|
|
|
case SamplerType.Texture2D:
|
|
return Target.Texture2D;
|
|
|
|
case SamplerType.Texture2D | SamplerType.Array:
|
|
return Target.Texture2DArray;
|
|
|
|
case SamplerType.Texture2D | SamplerType.Multisample:
|
|
return Target.Texture2DMultisample;
|
|
|
|
case SamplerType.Texture2D | SamplerType.Multisample | SamplerType.Array:
|
|
return Target.Texture2DMultisampleArray;
|
|
|
|
case SamplerType.Texture3D:
|
|
return Target.Texture3D;
|
|
|
|
case SamplerType.TextureCube:
|
|
return Target.Cubemap;
|
|
|
|
case SamplerType.TextureCube | SamplerType.Array:
|
|
return Target.CubemapArray;
|
|
}
|
|
|
|
Logger.Warning?.Print(LogClass.Gpu, $"Invalid sampler type \"{type}\".");
|
|
|
|
return Target.Texture2D;
|
|
}
|
|
|
|
/// <summary>
|
|
/// Issues a texture barrier.
|
|
/// This waits until previous texture writes from the GPU to finish, before
|
|
/// performing new operations with said textures.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state (unused)</param>
|
|
/// <param name="argument">Method call argument (unused)</param>
|
|
private void TextureBarrier(GpuState state, int argument)
|
|
{
|
|
_context.Renderer.Pipeline.TextureBarrier();
|
|
}
|
|
|
|
/// <summary>
|
|
/// Invalidates all modified textures on the cache.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state (unused)</param>
|
|
/// <param name="argument">Method call argument (unused)</param>
|
|
private void InvalidateTextures(GpuState state, int argument)
|
|
{
|
|
TextureManager.Flush();
|
|
}
|
|
|
|
/// <summary>
|
|
/// Issues a texture barrier.
|
|
/// This waits until previous texture writes from the GPU to finish, before
|
|
/// performing new operations with said textures.
|
|
/// This performs a per-tile wait, it is only valid if both the previous write
|
|
/// and current access has the same access patterns.
|
|
/// This may be faster than the regular barrier on tile-based rasterizers.
|
|
/// </summary>
|
|
/// <param name="state">Current GPU state (unused)</param>
|
|
/// <param name="argument">Method call argument (unused)</param>
|
|
private void TextureBarrierTiled(GpuState state, int argument)
|
|
{
|
|
_context.Renderer.Pipeline.TextureBarrierTiled();
|
|
}
|
|
}
|
|
} |