using Ryujinx.Graphics.GAL;
using Ryujinx.Memory.Range;
using System;
using System.Collections.Generic;
using System.Linq;
namespace Ryujinx.Graphics.Gpu.Memory
{
///
/// Buffer cache.
///
class BufferCache : IDisposable
{
private const int OverlapsBufferInitialCapacity = 10;
private const int OverlapsBufferMaxCapacity = 10000;
private const ulong BufferAlignmentSize = 0x1000;
private const ulong BufferAlignmentMask = BufferAlignmentSize - 1;
private readonly GpuContext _context;
private readonly PhysicalMemory _physicalMemory;
private readonly RangeList _buffers;
private Buffer[] _bufferOverlaps;
private readonly Dictionary _dirtyCache;
public event Action NotifyBuffersModified;
///
/// Creates a new instance of the buffer manager.
///
/// The GPU context that the buffer manager belongs to
/// Physical memory where the cached buffers are mapped
public BufferCache(GpuContext context, PhysicalMemory physicalMemory)
{
_context = context;
_physicalMemory = physicalMemory;
_buffers = new RangeList();
_bufferOverlaps = new Buffer[OverlapsBufferInitialCapacity];
_dirtyCache = new Dictionary();
}
///
/// Handles removal of buffers written to a memory region being unmapped.
///
/// Sender object
/// Event arguments
public void MemoryUnmappedHandler(object sender, UnmapEventArgs e)
{
Buffer[] overlaps = new Buffer[10];
int overlapCount;
ulong address = ((MemoryManager)sender).Translate(e.Address);
ulong size = e.Size;
lock (_buffers)
{
overlapCount = _buffers.FindOverlaps(address, size, ref overlaps);
}
for (int i = 0; i < overlapCount; i++)
{
overlaps[i].Unmapped(address, size);
}
}
///
/// Performs address translation of the GPU virtual address, and creates a
/// new buffer, if needed, for the specified range.
///
/// GPU memory manager where the buffer is mapped
/// Start GPU virtual address of the buffer
/// Size in bytes of the buffer
/// CPU virtual address of the buffer, after address translation
public ulong TranslateAndCreateBuffer(MemoryManager memoryManager, ulong gpuVa, ulong size)
{
if (gpuVa == 0)
{
return 0;
}
ulong address = memoryManager.Translate(gpuVa);
if (address == MemoryManager.PteUnmapped)
{
return 0;
}
CreateBuffer(address, size);
return address;
}
///
/// Creates a new buffer for the specified range, if it does not yet exist.
/// This can be used to ensure the existance of a buffer.
///
/// Address of the buffer in memory
/// Size of the buffer in bytes
public void CreateBuffer(ulong address, ulong size)
{
ulong endAddress = address + size;
ulong alignedAddress = address & ~BufferAlignmentMask;
ulong alignedEndAddress = (endAddress + BufferAlignmentMask) & ~BufferAlignmentMask;
// The buffer must have the size of at least one page.
if (alignedEndAddress == alignedAddress)
{
alignedEndAddress += BufferAlignmentSize;
}
CreateBufferAligned(alignedAddress, alignedEndAddress - alignedAddress);
}
///
/// Performs address translation of the GPU virtual address, and attempts to force
/// the buffer in the region as dirty.
/// The buffer lookup for this function is cached in a dictionary for quick access, which
/// accelerates common UBO updates.
///
/// GPU memory manager where the buffer is mapped
/// Start GPU virtual address of the buffer
/// Size in bytes of the buffer
public void ForceDirty(MemoryManager memoryManager, ulong gpuVa, ulong size)
{
if (!_dirtyCache.TryGetValue(gpuVa, out BufferCacheEntry result) ||
result.EndGpuAddress < gpuVa + size ||
result.UnmappedSequence != result.Buffer.UnmappedSequence)
{
ulong address = TranslateAndCreateBuffer(memoryManager, gpuVa, size);
result = new BufferCacheEntry(address, gpuVa, GetBuffer(address, size));
_dirtyCache[gpuVa] = result;
}
result.Buffer.ForceDirty(result.Address, size);
}
///
/// Creates a new buffer for the specified range, if needed.
/// If a buffer where this range can be fully contained already exists,
/// then the creation of a new buffer is not necessary.
///
/// Address of the buffer in guest memory
/// Size in bytes of the buffer
private void CreateBufferAligned(ulong address, ulong size)
{
int overlapsCount;
lock (_buffers)
{
overlapsCount = _buffers.FindOverlapsNonOverlapping(address, size, ref _bufferOverlaps);
}
if (overlapsCount != 0)
{
// The buffer already exists. We can just return the existing buffer
// if the buffer we need is fully contained inside the overlapping buffer.
// Otherwise, we must delete the overlapping buffers and create a bigger buffer
// that fits all the data we need. We also need to copy the contents from the
// old buffer(s) to the new buffer.
ulong endAddress = address + size;
if (_bufferOverlaps[0].Address > address || _bufferOverlaps[0].EndAddress < endAddress)
{
for (int index = 0; index < overlapsCount; index++)
{
Buffer buffer = _bufferOverlaps[index];
address = Math.Min(address, buffer.Address);
endAddress = Math.Max(endAddress, buffer.EndAddress);
lock (_buffers)
{
_buffers.Remove(buffer);
}
}
Buffer newBuffer = new Buffer(_context, _physicalMemory, address, endAddress - address, _bufferOverlaps.Take(overlapsCount));
lock (_buffers)
{
_buffers.Add(newBuffer);
}
for (int index = 0; index < overlapsCount; index++)
{
Buffer buffer = _bufferOverlaps[index];
int dstOffset = (int)(buffer.Address - newBuffer.Address);
buffer.CopyTo(newBuffer, dstOffset);
newBuffer.InheritModifiedRanges(buffer);
buffer.DisposeData();
}
newBuffer.SynchronizeMemory(address, endAddress - address);
// Existing buffers were modified, we need to rebind everything.
NotifyBuffersModified?.Invoke();
}
}
else
{
// No overlap, just create a new buffer.
Buffer buffer = new Buffer(_context, _physicalMemory, address, size);
lock (_buffers)
{
_buffers.Add(buffer);
}
}
ShrinkOverlapsBufferIfNeeded();
}
///
/// Resizes the temporary buffer used for range list intersection results, if it has grown too much.
///
private void ShrinkOverlapsBufferIfNeeded()
{
if (_bufferOverlaps.Length > OverlapsBufferMaxCapacity)
{
Array.Resize(ref _bufferOverlaps, OverlapsBufferMaxCapacity);
}
}
///
/// Copy a buffer data from a given address to another.
///
///
/// This does a GPU side copy.
///
/// GPU memory manager where the buffer is mapped
/// GPU virtual address of the copy source
/// GPU virtual address of the copy destination
/// Size in bytes of the copy
public void CopyBuffer(MemoryManager memoryManager, ulong srcVa, ulong dstVa, ulong size)
{
ulong srcAddress = TranslateAndCreateBuffer(memoryManager, srcVa, size);
ulong dstAddress = TranslateAndCreateBuffer(memoryManager, dstVa, size);
Buffer srcBuffer = GetBuffer(srcAddress, size);
Buffer dstBuffer = GetBuffer(dstAddress, size);
int srcOffset = (int)(srcAddress - srcBuffer.Address);
int dstOffset = (int)(dstAddress - dstBuffer.Address);
_context.Renderer.Pipeline.CopyBuffer(
srcBuffer.Handle,
dstBuffer.Handle,
srcOffset,
dstOffset,
(int)size);
if (srcBuffer.IsModified(srcAddress, size))
{
dstBuffer.SignalModified(dstAddress, size);
}
else
{
// Optimization: If the data being copied is already in memory, then copy it directly instead of flushing from GPU.
dstBuffer.ClearModified(dstAddress, size);
memoryManager.Physical.WriteUntracked(dstAddress, memoryManager.Physical.GetSpan(srcAddress, (int)size));
}
}
///
/// Clears a buffer at a given address with the specified value.
///
///
/// Both the address and size must be aligned to 4 bytes.
///
/// GPU memory manager where the buffer is mapped
/// GPU virtual address of the region to clear
/// Number of bytes to clear
/// Value to be written into the buffer
public void ClearBuffer(MemoryManager memoryManager, ulong gpuVa, ulong size, uint value)
{
ulong address = TranslateAndCreateBuffer(memoryManager, gpuVa, size);
Buffer buffer = GetBuffer(address, size);
int offset = (int)(address - buffer.Address);
_context.Renderer.Pipeline.ClearBuffer(buffer.Handle, offset, (int)size, value);
buffer.SignalModified(address, size);
}
///
/// Gets a buffer sub-range starting at a given memory address.
///
/// Start address of the memory range
/// Size in bytes of the memory range
/// Whether the buffer will be written to by this use
/// The buffer sub-range starting at the given memory address
public BufferRange GetBufferRangeTillEnd(ulong address, ulong size, bool write = false)
{
return GetBuffer(address, size, write).GetRange(address);
}
///
/// Gets a buffer sub-range for a given memory range.
///
/// Start address of the memory range
/// Size in bytes of the memory range
/// Whether the buffer will be written to by this use
/// The buffer sub-range for the given range
public BufferRange GetBufferRange(ulong address, ulong size, bool write = false)
{
return GetBuffer(address, size, write).GetRange(address, size);
}
///
/// Gets a buffer for a given memory range.
/// A buffer overlapping with the specified range is assumed to already exist on the cache.
///
/// Start address of the memory range
/// Size in bytes of the memory range
/// Whether the buffer will be written to by this use
/// The buffer where the range is fully contained
private Buffer GetBuffer(ulong address, ulong size, bool write = false)
{
Buffer buffer;
if (size != 0)
{
lock (_buffers)
{
buffer = _buffers.FindFirstOverlap(address, size);
}
buffer.SynchronizeMemory(address, size);
if (write)
{
buffer.SignalModified(address, size);
}
}
else
{
lock (_buffers)
{
buffer = _buffers.FindFirstOverlap(address, 1);
}
}
return buffer;
}
///
/// Performs guest to host memory synchronization of a given memory range.
///
/// Start address of the memory range
/// Size in bytes of the memory range
public void SynchronizeBufferRange(ulong address, ulong size)
{
if (size != 0)
{
Buffer buffer;
lock (_buffers)
{
buffer = _buffers.FindFirstOverlap(address, size);
}
buffer.SynchronizeMemory(address, size);
}
}
///
/// Disposes all buffers in the cache.
/// It's an error to use the buffer manager after disposal.
///
public void Dispose()
{
lock (_buffers)
{
foreach (Buffer buffer in _buffers)
{
buffer.Dispose();
}
}
}
}
}