pcsx2/plugins/GSdx/GSLocalMemory.cpp

/*
 *	Copyright (C) 2007-2009 Gabest
 *	http://www.gabest.org
 *
 *  This Program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2, or (at your option)
 *  any later version.
 *
 *  This Program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with GNU Make; see the file COPYING.  If not, write to
 *  the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA USA.
 *  http://www.gnu.org/copyleft/gpl.html
 *
 *	Special Notes:
 *
 *	Based on Page.c from GSSoft
 *	Copyright (C) 2002-2004 GSsoft Team
 *
 */

#include "stdafx.h"
#include "GSLocalMemory.h"
#include "GSdx.h"

#define ASSERT_BLOCK(r, w, h) \
	ASSERT((r).width() >= (w) && (r).height() >= (h) && !((r).left & ((w) - 1)) && !((r).top & ((h) - 1)) && !((r).right & ((w) - 1)) && !((r).bottom & ((h) - 1))); \

#define FOREACH_BLOCK_START(r, w, h, bpp) \
	ASSERT_BLOCK(r, w, h); \
	GSVector4i _r = (r) >> 3; \
	uint8* _dst = dst - _r.left * (bpp); \
	int _offset = dstpitch * (h); \
	for(int y = _r.top; y < _r.bottom; y += (h) >> 3, _dst += _offset) \
	{ \
		uint32 _base = off->block.row[y]; \
		for(int x = _r.left; x < _r.right; x += (w) >> 3) \
		{ \
			const uint8* src = BlockPtr(_base + off->block.col[x]); \
			uint8* read_dst = &_dst[x * (bpp)]; \

#define FOREACH_BLOCK_END }}

//

uint32 GSLocalMemory::pageOffset32[32][32][64];
uint32 GSLocalMemory::pageOffset32Z[32][32][64];
uint32 GSLocalMemory::pageOffset16[32][64][64];
uint32 GSLocalMemory::pageOffset16S[32][64][64];
uint32 GSLocalMemory::pageOffset16Z[32][64][64];
uint32 GSLocalMemory::pageOffset16SZ[32][64][64];
uint32 GSLocalMemory::pageOffset8[32][64][128];
uint32 GSLocalMemory::pageOffset4[32][128][128];

int GSLocalMemory::rowOffset32[4096];
int GSLocalMemory::rowOffset32Z[4096];
int GSLocalMemory::rowOffset16[4096];
int GSLocalMemory::rowOffset16S[4096];
int GSLocalMemory::rowOffset16Z[4096];
int GSLocalMemory::rowOffset16SZ[4096];
int GSLocalMemory::rowOffset8[2][4096];
int GSLocalMemory::rowOffset4[2][4096];

short GSLocalMemory::blockOffset32[256];
short GSLocalMemory::blockOffset32Z[256];
short GSLocalMemory::blockOffset16[256];
short GSLocalMemory::blockOffset16S[256];
short GSLocalMemory::blockOffset16Z[256];
short GSLocalMemory::blockOffset16SZ[256];
short GSLocalMemory::blockOffset8[256];
short GSLocalMemory::blockOffset4[256];

//

GSLocalMemory::psm_t GSLocalMemory::m_psm[64];

//

GSLocalMemory::GSLocalMemory()
	: m_clut(this)
{
	m_use_fifo_alloc = theApp.GetConfigB("UserHacks") && theApp.GetConfigB("wrap_gs_mem");
	switch (theApp.GetCurrentRendererType()) {
		case GSRendererType::OGL_SW:
		case GSRendererType::DX9_SW:
		case GSRendererType::DX1011_SW:
			m_use_fifo_alloc = true;
			break;
		default: break;
	}

	if (m_use_fifo_alloc)
		m_vm8 = (uint8*)fifo_alloc(m_vmsize, 4);
	else
		m_vm8 = nullptr;

	// Either we don't use fifo alloc or we get an error.
	if (m_vm8 == nullptr)
	{
		m_vm8 = (uint8*)vmalloc(m_vmsize * 4, false);
		m_use_fifo_alloc = false;
	}

	m_vm16 = (uint16*)m_vm8;
	m_vm32 = (uint32*)m_vm8;

	memset(m_vm8, 0, m_vmsize);

	for(int bp = 0; bp < 32; bp++)
	{
		for(int y = 0; y < 32; y++) for(int x = 0; x < 64; x++)
		{
			pageOffset32[bp][y][x] = PixelAddressOrg32(x, y, bp, 0);
			pageOffset32Z[bp][y][x] = PixelAddressOrg32Z(x, y, bp, 0);
		}

		for(int y = 0; y < 64; y++) for(int x = 0; x < 64; x++)
		{
			pageOffset16[bp][y][x] = PixelAddressOrg16(x, y, bp, 0);
			pageOffset16S[bp][y][x] = PixelAddressOrg16S(x, y, bp, 0);
			pageOffset16Z[bp][y][x] = PixelAddressOrg16Z(x, y, bp, 0);
			pageOffset16SZ[bp][y][x] = PixelAddressOrg16SZ(x, y, bp, 0);
		}

		for(int y = 0; y < 64; y++) for(int x = 0; x < 128; x++)
		{
			pageOffset8[bp][y][x] = PixelAddressOrg8(x, y, bp, 0);
		}

		for(int y = 0; y < 128; y++) for(int x = 0; x < 128; x++)
		{
			pageOffset4[bp][y][x] = PixelAddressOrg4(x, y, bp, 0);
		}
	}

	for(size_t x = 0; x < countof(rowOffset32); x++)
	{
		rowOffset32[x] = (int)PixelAddress32(x & 0x7ff, 0, 0, 32) - (int)PixelAddress32(0, 0, 0, 32);
	}

	for(size_t x = 0; x < countof(rowOffset32Z); x++)
	{
		rowOffset32Z[x] = (int)PixelAddress32Z(x & 0x7ff, 0, 0, 32) - (int)PixelAddress32Z(0, 0, 0, 32);
	}

	for(size_t x = 0; x < countof(rowOffset16); x++)
	{
		rowOffset16[x] = (int)PixelAddress16(x & 0x7ff, 0, 0, 32) - (int)PixelAddress16(0, 0, 0, 32);
	}

	for(size_t x = 0; x < countof(rowOffset16S); x++)
	{
		rowOffset16S[x] = (int)PixelAddress16S(x & 0x7ff, 0, 0, 32) - (int)PixelAddress16S(0, 0, 0, 32);
	}

	for(size_t x = 0; x < countof(rowOffset16Z); x++)
	{
		rowOffset16Z[x] = (int)PixelAddress16Z(x & 0x7ff, 0, 0, 32) - (int)PixelAddress16Z(0, 0, 0, 32);
	}

	for(size_t x = 0; x < countof(rowOffset16SZ); x++)
	{
		rowOffset16SZ[x] = (int)PixelAddress16SZ(x & 0x7ff, 0, 0, 32) - (int)PixelAddress16SZ(0, 0, 0, 32);
	}

	for(size_t x = 0; x < countof(rowOffset8[0]); x++)
	{
		rowOffset8[0][x] = (int)PixelAddress8(x & 0x7ff, 0, 0, 32) - (int)PixelAddress8(0, 0, 0, 32);
		rowOffset8[1][x] = (int)PixelAddress8(x & 0x7ff, 2, 0, 32) - (int)PixelAddress8(0, 2, 0, 32);
	}

	for(size_t x = 0; x < countof(rowOffset4[0]); x++)
	{
		rowOffset4[0][x] = (int)PixelAddress4(x & 0x7ff, 0, 0, 32) - (int)PixelAddress4(0, 0, 0, 32);
		rowOffset4[1][x] = (int)PixelAddress4(x & 0x7ff, 2, 0, 32) - (int)PixelAddress4(0, 2, 0, 32);
	}

	for(size_t x = 0; x < countof(blockOffset32); x++)
	{
		blockOffset32[x] = (short)((int)BlockNumber32(x << 3, 0, 0, 32) - (int)BlockNumber32(0, 0, 0, 32));
	}

	for(size_t x = 0; x < countof(blockOffset32Z); x++)
	{
		blockOffset32Z[x] = (short)((int)BlockNumber32Z(x << 3, 0, 0, 32) - (int)BlockNumber32Z(0, 0, 0, 32));
	}

	for(size_t x = 0; x < countof(blockOffset16); x++)
	{
		blockOffset16[x] = (short)((int)BlockNumber16(x << 3, 0, 0, 32) - (int)BlockNumber16(0, 0, 0, 32));
	}

	for(size_t x = 0; x < countof(blockOffset16S); x++)
	{
		blockOffset16S[x] = (short)((int)BlockNumber16S(x << 3, 0, 0, 32) - (int)BlockNumber16S(0, 0, 0, 32));
	}

	for(size_t x = 0; x < countof(blockOffset16Z); x++)
	{
		blockOffset16Z[x] = (short)((int)BlockNumber16Z(x << 3, 0, 0, 32) - (int)BlockNumber16Z(0, 0, 0, 32));
	}

	for(size_t x = 0; x < countof(blockOffset16SZ); x++)
	{
		blockOffset16SZ[x] = (short)((int)BlockNumber16SZ(x << 3, 0, 0, 32) - (int)BlockNumber16SZ(0, 0, 0, 32));
	}

	for(size_t x = 0; x < countof(blockOffset8); x++)
	{
		blockOffset8[x] = (short)((int)BlockNumber8(x << 3, 0, 0, 32) - (int)BlockNumber8(0, 0, 0, 32));
	}

	for(size_t x = 0; x < countof(blockOffset4); x++)
	{
		blockOffset4[x] = (short)((int)BlockNumber4(x << 3, 0, 0, 32) - (int)BlockNumber4(0, 0, 0, 32));
	}

	for(size_t i = 0; i < countof(m_psm); i++)
	{
		m_psm[i].pa = &GSLocalMemory::PixelAddress32;
		m_psm[i].bn = &GSLocalMemory::BlockNumber32;
		m_psm[i].rp = &GSLocalMemory::ReadPixel32;
		m_psm[i].rpa = &GSLocalMemory::ReadPixel32;
		m_psm[i].wp = &GSLocalMemory::WritePixel32;
		m_psm[i].wpa = &GSLocalMemory::WritePixel32;
		m_psm[i].rt = &GSLocalMemory::ReadTexel32;
		m_psm[i].rta = &GSLocalMemory::ReadTexel32;
		m_psm[i].wfa = &GSLocalMemory::WritePixel32;
		m_psm[i].wi = &GSLocalMemory::WriteImage<PSM_PSMCT32, 8, 8, 32>;
		m_psm[i].ri = &GSLocalMemory::ReadImageX; // TODO
		m_psm[i].rtx = &GSLocalMemory::ReadTexture32;
		m_psm[i].rtxP = &GSLocalMemory::ReadTexture32;
		m_psm[i].rtxb = &GSLocalMemory::ReadTextureBlock32;
		m_psm[i].rtxbP = &GSLocalMemory::ReadTextureBlock32;
		m_psm[i].bpp = m_psm[i].trbpp = 32;
		m_psm[i].pal = 0;
		m_psm[i].bs = GSVector2i(8, 8);
		m_psm[i].pgs = GSVector2i(64, 32);
		for(int j = 0; j < 8; j++) m_psm[i].rowOffset[j] = rowOffset32;
		m_psm[i].blockOffset = blockOffset32;
		m_psm[i].msk = 0xff;
		m_psm[i].depth = 0;
	}

	m_psm[PSM_PSGPU24].pa = &GSLocalMemory::PixelAddress16;
	m_psm[PSM_PSMCT16].pa = &GSLocalMemory::PixelAddress16;
	m_psm[PSM_PSMCT16S].pa = &GSLocalMemory::PixelAddress16S;
	m_psm[PSM_PSMT8].pa = &GSLocalMemory::PixelAddress8;
	m_psm[PSM_PSMT4].pa = &GSLocalMemory::PixelAddress4;
	m_psm[PSM_PSMZ32].pa = &GSLocalMemory::PixelAddress32Z;
	m_psm[PSM_PSMZ24].pa = &GSLocalMemory::PixelAddress32Z;
	m_psm[PSM_PSMZ16].pa = &GSLocalMemory::PixelAddress16Z;
	m_psm[PSM_PSMZ16S].pa = &GSLocalMemory::PixelAddress16SZ;

	m_psm[PSM_PSGPU24].bn = &GSLocalMemory::BlockNumber16;
	m_psm[PSM_PSMCT16].bn = &GSLocalMemory::BlockNumber16;
	m_psm[PSM_PSMCT16S].bn = &GSLocalMemory::BlockNumber16S;
	m_psm[PSM_PSMT8].bn = &GSLocalMemory::BlockNumber8;
	m_psm[PSM_PSMT4].bn = &GSLocalMemory::BlockNumber4;
	m_psm[PSM_PSMZ32].bn = &GSLocalMemory::BlockNumber32Z;
	m_psm[PSM_PSMZ24].bn = &GSLocalMemory::BlockNumber32Z;
	m_psm[PSM_PSMZ16].bn = &GSLocalMemory::BlockNumber16Z;
	m_psm[PSM_PSMZ16S].bn = &GSLocalMemory::BlockNumber16SZ;

	m_psm[PSM_PSMCT24].rp = &GSLocalMemory::ReadPixel24;
	m_psm[PSM_PSMCT16].rp = &GSLocalMemory::ReadPixel16;
	m_psm[PSM_PSMCT16S].rp = &GSLocalMemory::ReadPixel16S;
	m_psm[PSM_PSMT8].rp = &GSLocalMemory::ReadPixel8;
	m_psm[PSM_PSMT4].rp = &GSLocalMemory::ReadPixel4;
	m_psm[PSM_PSMT8H].rp = &GSLocalMemory::ReadPixel8H;
	m_psm[PSM_PSMT4HL].rp = &GSLocalMemory::ReadPixel4HL;
	m_psm[PSM_PSMT4HH].rp = &GSLocalMemory::ReadPixel4HH;
	m_psm[PSM_PSMZ32].rp = &GSLocalMemory::ReadPixel32Z;
	m_psm[PSM_PSMZ24].rp = &GSLocalMemory::ReadPixel24Z;
	m_psm[PSM_PSMZ16].rp = &GSLocalMemory::ReadPixel16Z;
	m_psm[PSM_PSMZ16S].rp = &GSLocalMemory::ReadPixel16SZ;

	m_psm[PSM_PSMCT24].rpa = &GSLocalMemory::ReadPixel24;
	m_psm[PSM_PSMCT16].rpa = &GSLocalMemory::ReadPixel16;
	m_psm[PSM_PSMCT16S].rpa = &GSLocalMemory::ReadPixel16;
	m_psm[PSM_PSMT8].rpa = &GSLocalMemory::ReadPixel8;
	m_psm[PSM_PSMT4].rpa = &GSLocalMemory::ReadPixel4;
	m_psm[PSM_PSMT8H].rpa = &GSLocalMemory::ReadPixel8H;
	m_psm[PSM_PSMT4HL].rpa = &GSLocalMemory::ReadPixel4HL;
	m_psm[PSM_PSMT4HH].rpa = &GSLocalMemory::ReadPixel4HH;
	m_psm[PSM_PSMZ32].rpa = &GSLocalMemory::ReadPixel32;
	m_psm[PSM_PSMZ24].rpa = &GSLocalMemory::ReadPixel24;
	m_psm[PSM_PSMZ16].rpa = &GSLocalMemory::ReadPixel16;
	m_psm[PSM_PSMZ16S].rpa = &GSLocalMemory::ReadPixel16;

	m_psm[PSM_PSMCT32].wp = &GSLocalMemory::WritePixel32;
	m_psm[PSM_PSMCT24].wp = &GSLocalMemory::WritePixel24;
	m_psm[PSM_PSMCT16].wp = &GSLocalMemory::WritePixel16;
	m_psm[PSM_PSMCT16S].wp = &GSLocalMemory::WritePixel16S;
	m_psm[PSM_PSMT8].wp = &GSLocalMemory::WritePixel8;
	m_psm[PSM_PSMT4].wp = &GSLocalMemory::WritePixel4;
	m_psm[PSM_PSMT8H].wp = &GSLocalMemory::WritePixel8H;
	m_psm[PSM_PSMT4HL].wp = &GSLocalMemory::WritePixel4HL;
	m_psm[PSM_PSMT4HH].wp = &GSLocalMemory::WritePixel4HH;
	m_psm[PSM_PSMZ32].wp = &GSLocalMemory::WritePixel32Z;
	m_psm[PSM_PSMZ24].wp = &GSLocalMemory::WritePixel24Z;
	m_psm[PSM_PSMZ16].wp = &GSLocalMemory::WritePixel16Z;
	m_psm[PSM_PSMZ16S].wp = &GSLocalMemory::WritePixel16SZ;

	m_psm[PSM_PSMCT32].wpa = &GSLocalMemory::WritePixel32;
	m_psm[PSM_PSMCT24].wpa = &GSLocalMemory::WritePixel24;
	m_psm[PSM_PSMCT16].wpa = &GSLocalMemory::WritePixel16;
	m_psm[PSM_PSMCT16S].wpa = &GSLocalMemory::WritePixel16;
	m_psm[PSM_PSMT8].wpa = &GSLocalMemory::WritePixel8;
	m_psm[PSM_PSMT4].wpa = &GSLocalMemory::WritePixel4;
	m_psm[PSM_PSMT8H].wpa = &GSLocalMemory::WritePixel8H;
	m_psm[PSM_PSMT4HL].wpa = &GSLocalMemory::WritePixel4HL;
	m_psm[PSM_PSMT4HH].wpa = &GSLocalMemory::WritePixel4HH;
	m_psm[PSM_PSMZ32].wpa = &GSLocalMemory::WritePixel32;
	m_psm[PSM_PSMZ24].wpa = &GSLocalMemory::WritePixel24;
	m_psm[PSM_PSMZ16].wpa = &GSLocalMemory::WritePixel16;
	m_psm[PSM_PSMZ16S].wpa = &GSLocalMemory::WritePixel16;

	m_psm[PSM_PSMCT24].rt = &GSLocalMemory::ReadTexel24;
	m_psm[PSM_PSMCT16].rt = &GSLocalMemory::ReadTexel16;
	m_psm[PSM_PSMCT16S].rt = &GSLocalMemory::ReadTexel16S;
	m_psm[PSM_PSMT8].rt = &GSLocalMemory::ReadTexel8;
	m_psm[PSM_PSMT4].rt = &GSLocalMemory::ReadTexel4;
	m_psm[PSM_PSMT8H].rt = &GSLocalMemory::ReadTexel8H;
	m_psm[PSM_PSMT4HL].rt = &GSLocalMemory::ReadTexel4HL;
	m_psm[PSM_PSMT4HH].rt = &GSLocalMemory::ReadTexel4HH;
	m_psm[PSM_PSMZ32].rt = &GSLocalMemory::ReadTexel32Z;
	m_psm[PSM_PSMZ24].rt = &GSLocalMemory::ReadTexel24Z;
	m_psm[PSM_PSMZ16].rt = &GSLocalMemory::ReadTexel16Z;
	m_psm[PSM_PSMZ16S].rt = &GSLocalMemory::ReadTexel16SZ;

	m_psm[PSM_PSMCT24].rta = &GSLocalMemory::ReadTexel24;
	m_psm[PSM_PSMCT16].rta = &GSLocalMemory::ReadTexel16;
	m_psm[PSM_PSMCT16S].rta = &GSLocalMemory::ReadTexel16;
	m_psm[PSM_PSMT8].rta = &GSLocalMemory::ReadTexel8;
	m_psm[PSM_PSMT4].rta = &GSLocalMemory::ReadTexel4;
	m_psm[PSM_PSMT8H].rta = &GSLocalMemory::ReadTexel8H;
	m_psm[PSM_PSMT4HL].rta = &GSLocalMemory::ReadTexel4HL;
	m_psm[PSM_PSMT4HH].rta = &GSLocalMemory::ReadTexel4HH;
	m_psm[PSM_PSMZ24].rta = &GSLocalMemory::ReadTexel24;
	m_psm[PSM_PSMZ16].rta = &GSLocalMemory::ReadTexel16;
	m_psm[PSM_PSMZ16S].rta = &GSLocalMemory::ReadTexel16;

	m_psm[PSM_PSMCT24].wfa = &GSLocalMemory::WritePixel24;
	m_psm[PSM_PSMCT16].wfa = &GSLocalMemory::WriteFrame16;
	m_psm[PSM_PSMCT16S].wfa = &GSLocalMemory::WriteFrame16;
	m_psm[PSM_PSMZ24].wfa = &GSLocalMemory::WritePixel24;
	m_psm[PSM_PSMZ16].wfa = &GSLocalMemory::WriteFrame16;
	m_psm[PSM_PSMZ16S].wfa = &GSLocalMemory::WriteFrame16;

	m_psm[PSM_PSMCT24].wi = &GSLocalMemory::WriteImage24; // TODO
	m_psm[PSM_PSMCT16].wi = &GSLocalMemory::WriteImage<PSM_PSMCT16, 16, 8, 16>;
	m_psm[PSM_PSMCT16S].wi = &GSLocalMemory::WriteImage<PSM_PSMCT16S, 16, 8, 16>;
	m_psm[PSM_PSMT8].wi = &GSLocalMemory::WriteImage<PSM_PSMT8, 16, 16, 8>;
	m_psm[PSM_PSMT4].wi = &GSLocalMemory::WriteImage<PSM_PSMT4, 32, 16, 4>;
	m_psm[PSM_PSMT8H].wi = &GSLocalMemory::WriteImage8H; // TODO
	m_psm[PSM_PSMT4HL].wi = &GSLocalMemory::WriteImage4HL; // TODO
	m_psm[PSM_PSMT4HH].wi = &GSLocalMemory::WriteImage4HH; // TODO
	m_psm[PSM_PSMZ32].wi = &GSLocalMemory::WriteImage<PSM_PSMZ32, 8, 8, 32>;
	m_psm[PSM_PSMZ24].wi = &GSLocalMemory::WriteImage24Z; // TODO
	m_psm[PSM_PSMZ16].wi = &GSLocalMemory::WriteImage<PSM_PSMZ16, 16, 8, 16>;
	m_psm[PSM_PSMZ16S].wi = &GSLocalMemory::WriteImage<PSM_PSMZ16S, 16, 8, 16>;

	m_psm[PSM_PSMCT24].rtx = &GSLocalMemory::ReadTexture24;
	m_psm[PSM_PSGPU24].rtx = &GSLocalMemory::ReadTextureGPU24;
	m_psm[PSM_PSMCT16].rtx = &GSLocalMemory::ReadTexture16;
	m_psm[PSM_PSMCT16S].rtx = &GSLocalMemory::ReadTexture16;
	m_psm[PSM_PSMT8].rtx = &GSLocalMemory::ReadTexture8;
	m_psm[PSM_PSMT4].rtx = &GSLocalMemory::ReadTexture4;
	m_psm[PSM_PSMT8H].rtx = &GSLocalMemory::ReadTexture8H;
	m_psm[PSM_PSMT4HL].rtx = &GSLocalMemory::ReadTexture4HL;
	m_psm[PSM_PSMT4HH].rtx = &GSLocalMemory::ReadTexture4HH;
	m_psm[PSM_PSMZ32].rtx = &GSLocalMemory::ReadTexture32;
	m_psm[PSM_PSMZ24].rtx = &GSLocalMemory::ReadTexture24;
	m_psm[PSM_PSMZ16].rtx = &GSLocalMemory::ReadTexture16;
	m_psm[PSM_PSMZ16S].rtx = &GSLocalMemory::ReadTexture16;

	m_psm[PSM_PSMCT24].rtxP = &GSLocalMemory::ReadTexture24;
	m_psm[PSM_PSMCT16].rtxP = &GSLocalMemory::ReadTexture16;
	m_psm[PSM_PSMCT16S].rtxP = &GSLocalMemory::ReadTexture16;
	m_psm[PSM_PSMT8].rtxP = &GSLocalMemory::ReadTexture8P;
	m_psm[PSM_PSMT4].rtxP = &GSLocalMemory::ReadTexture4P;
	m_psm[PSM_PSMT8H].rtxP = &GSLocalMemory::ReadTexture8HP;
	m_psm[PSM_PSMT4HL].rtxP = &GSLocalMemory::ReadTexture4HLP;
	m_psm[PSM_PSMT4HH].rtxP = &GSLocalMemory::ReadTexture4HHP;
	m_psm[PSM_PSMZ32].rtxP = &GSLocalMemory::ReadTexture32;
	m_psm[PSM_PSMZ24].rtxP = &GSLocalMemory::ReadTexture24;
	m_psm[PSM_PSMZ16].rtxP = &GSLocalMemory::ReadTexture16;
	m_psm[PSM_PSMZ16S].rtxP = &GSLocalMemory::ReadTexture16;

	m_psm[PSM_PSMCT24].rtxb = &GSLocalMemory::ReadTextureBlock24;
	m_psm[PSM_PSMCT16].rtxb = &GSLocalMemory::ReadTextureBlock16;
	m_psm[PSM_PSMCT16S].rtxb = &GSLocalMemory::ReadTextureBlock16;
	m_psm[PSM_PSMT8].rtxb = &GSLocalMemory::ReadTextureBlock8;
	m_psm[PSM_PSMT4].rtxb = &GSLocalMemory::ReadTextureBlock4;
	m_psm[PSM_PSMT8H].rtxb = &GSLocalMemory::ReadTextureBlock8H;
	m_psm[PSM_PSMT4HL].rtxb = &GSLocalMemory::ReadTextureBlock4HL;
	m_psm[PSM_PSMT4HH].rtxb = &GSLocalMemory::ReadTextureBlock4HH;
	m_psm[PSM_PSMZ32].rtxb = &GSLocalMemory::ReadTextureBlock32;
	m_psm[PSM_PSMZ24].rtxb = &GSLocalMemory::ReadTextureBlock24;
	m_psm[PSM_PSMZ16].rtxb = &GSLocalMemory::ReadTextureBlock16;
	m_psm[PSM_PSMZ16S].rtxb = &GSLocalMemory::ReadTextureBlock16;

	m_psm[PSM_PSMCT24].rtxbP = &GSLocalMemory::ReadTextureBlock24;
	m_psm[PSM_PSMCT16].rtxbP = &GSLocalMemory::ReadTextureBlock16;
	m_psm[PSM_PSMCT16S].rtxbP = &GSLocalMemory::ReadTextureBlock16;
	m_psm[PSM_PSMT8].rtxbP = &GSLocalMemory::ReadTextureBlock8P;
	m_psm[PSM_PSMT4].rtxbP = &GSLocalMemory::ReadTextureBlock4P;
	m_psm[PSM_PSMT8H].rtxbP = &GSLocalMemory::ReadTextureBlock8HP;
	m_psm[PSM_PSMT4HL].rtxbP = &GSLocalMemory::ReadTextureBlock4HLP;
	m_psm[PSM_PSMT4HH].rtxbP = &GSLocalMemory::ReadTextureBlock4HHP;
	m_psm[PSM_PSMZ32].rtxbP = &GSLocalMemory::ReadTextureBlock32;
	m_psm[PSM_PSMZ24].rtxbP = &GSLocalMemory::ReadTextureBlock24;
	m_psm[PSM_PSMZ16].rtxbP = &GSLocalMemory::ReadTextureBlock16;
	m_psm[PSM_PSMZ16S].rtxbP = &GSLocalMemory::ReadTextureBlock16;

	m_psm[PSM_PSGPU24].bpp = 16;
	m_psm[PSM_PSMCT16].bpp = m_psm[PSM_PSMCT16S].bpp = 16;
	m_psm[PSM_PSMT8].bpp = 8;
	m_psm[PSM_PSMT4].bpp = 4;
	m_psm[PSM_PSMZ16].bpp = m_psm[PSM_PSMZ16S].bpp = 16;

	m_psm[PSM_PSMCT24].trbpp = 24;
	m_psm[PSM_PSGPU24].trbpp = 16;
	m_psm[PSM_PSMCT16].trbpp = m_psm[PSM_PSMCT16S].trbpp = 16;
	m_psm[PSM_PSMT8].trbpp = m_psm[PSM_PSMT8H].trbpp = 8;
	m_psm[PSM_PSMT4].trbpp = m_psm[PSM_PSMT4HL].trbpp = m_psm[PSM_PSMT4HH].trbpp = 4;
	m_psm[PSM_PSMZ24].trbpp = 24;
	m_psm[PSM_PSMZ16].trbpp = m_psm[PSM_PSMZ16S].trbpp = 16;

	m_psm[PSM_PSMT8].pal = m_psm[PSM_PSMT8H].pal = 256;
	m_psm[PSM_PSMT4].pal = m_psm[PSM_PSMT4HL].pal = m_psm[PSM_PSMT4HH].pal = 16;

	for(size_t i = 0; i < countof(m_psm); i++) m_psm[i].fmt = 3;
	m_psm[PSM_PSMCT32].fmt = m_psm[PSM_PSMZ32].fmt = 0;
	m_psm[PSM_PSMCT24].fmt = m_psm[PSM_PSMZ24].fmt = 1;
	m_psm[PSM_PSMCT16].fmt = m_psm[PSM_PSMZ16].fmt = 2;
	m_psm[PSM_PSMCT16S].fmt = m_psm[PSM_PSMZ16S].fmt = 2;


	m_psm[PSM_PSGPU24].bs = GSVector2i(16, 8);
	m_psm[PSM_PSMCT16].bs = m_psm[PSM_PSMCT16S].bs = GSVector2i(16, 8);
	m_psm[PSM_PSMT8].bs = GSVector2i(16, 16);
	m_psm[PSM_PSMT4].bs = GSVector2i(32, 16);
	m_psm[PSM_PSMZ16].bs = m_psm[PSM_PSMZ16S].bs = GSVector2i(16, 8);

	m_psm[PSM_PSGPU24].pgs = GSVector2i(64, 64);
	m_psm[PSM_PSMCT16].pgs = m_psm[PSM_PSMCT16S].pgs = GSVector2i(64, 64);
	m_psm[PSM_PSMT8].pgs = GSVector2i(128, 64);
	m_psm[PSM_PSMT4].pgs = GSVector2i(128, 128);
	m_psm[PSM_PSMZ16].pgs = m_psm[PSM_PSMZ16S].pgs = GSVector2i(64, 64);

	for(int i = 0; i < 8; i++) m_psm[PSM_PSGPU24].rowOffset[i] = rowOffset16;
	for(int i = 0; i < 8; i++) m_psm[PSM_PSMCT16].rowOffset[i] = rowOffset16;
	for(int i = 0; i < 8; i++) m_psm[PSM_PSMCT16S].rowOffset[i] = rowOffset16S;
	for(int i = 0; i < 8; i++) m_psm[PSM_PSMT8].rowOffset[i] = rowOffset8[((i + 2) >> 2) & 1];
	for(int i = 0; i < 8; i++) m_psm[PSM_PSMT4].rowOffset[i] = rowOffset4[((i + 2) >> 2) & 1];
	for(int i = 0; i < 8; i++) m_psm[PSM_PSMZ32].rowOffset[i] = rowOffset32Z;
	for(int i = 0; i < 8; i++) m_psm[PSM_PSMZ24].rowOffset[i] = rowOffset32Z;
	for(int i = 0; i < 8; i++) m_psm[PSM_PSMZ16].rowOffset[i] = rowOffset16Z;
	for(int i = 0; i < 8; i++) m_psm[PSM_PSMZ16S].rowOffset[i] = rowOffset16SZ;

	m_psm[PSM_PSGPU24].blockOffset = blockOffset16;
	m_psm[PSM_PSMCT16].blockOffset = blockOffset16;
	m_psm[PSM_PSMCT16S].blockOffset = blockOffset16S;
	m_psm[PSM_PSMT8].blockOffset = blockOffset8;
	m_psm[PSM_PSMT4].blockOffset = blockOffset4;
	m_psm[PSM_PSMZ32].blockOffset = blockOffset32Z;
	m_psm[PSM_PSMZ24].blockOffset = blockOffset32Z;
	m_psm[PSM_PSMZ16].blockOffset = blockOffset16Z;
	m_psm[PSM_PSMZ16S].blockOffset = blockOffset16SZ;

	m_psm[PSM_PSMCT24].msk = 0x3f;
	m_psm[PSM_PSMZ24].msk = 0x3f;
	m_psm[PSM_PSMT8H].msk = 0xc0;
	m_psm[PSM_PSMT4HL].msk = 0x40;
	m_psm[PSM_PSMT4HH].msk = 0x80;

	m_psm[PSM_PSMZ32].depth  = 1;
	m_psm[PSM_PSMZ24].depth  = 1;
	m_psm[PSM_PSMZ16].depth  = 1;
	m_psm[PSM_PSMZ16S].depth = 1;
}

GSLocalMemory::~GSLocalMemory()
{
	if (m_use_fifo_alloc)
		fifo_free(m_vm8, m_vmsize, 4);
	else
		vmfree(m_vm8, m_vmsize * 4);

	for(auto &i : m_omap) delete i.second;
	for(auto &i : m_pomap) _aligned_free(i.second);
	for(auto &i : m_po4map) _aligned_free(i.second);

	for(auto &i : m_p2tmap)
	{
		delete [] i.second;
	}
}

GSOffset* GSLocalMemory::GetOffset(uint32 bp, uint32 bw, uint32 psm)
{
	uint32 hash = bp | (bw << 14) | (psm << 20);

	auto i = m_omap.find(hash);

	if(i != m_omap.end())
	{
		return i->second;
	}

	GSOffset* off = new GSOffset(bp, bw, psm);

	m_omap[hash] = off;

	return off;
}

GSPixelOffset* GSLocalMemory::GetPixelOffset(const GIFRegFRAME& FRAME, const GIFRegZBUF& ZBUF)
{
	uint32 fbp = FRAME.Block();
	uint32 zbp = ZBUF.Block();
	uint32 fpsm = FRAME.PSM;
	uint32 zpsm = ZBUF.PSM;
	uint32 bw = FRAME.FBW;

	ASSERT(m_psm[fpsm].trbpp > 8 || m_psm[zpsm].trbpp > 8);

	// "(psm & 0x0f) ^ ((psm & 0xf0) >> 2)" creates 4 bit unique identifiers for render target formats (only)

	uint32 fpsm_hash = (fpsm & 0x0f) ^ ((fpsm & 0x30) >> 2);
	uint32 zpsm_hash = (zpsm & 0x0f) ^ ((zpsm & 0x30) >> 2);

	uint32 hash = (FRAME.FBP << 0) | (ZBUF.ZBP << 9) | (bw << 18) | (fpsm_hash << 24) | (zpsm_hash << 28);

	auto it = m_pomap.find(hash);

	if(it != m_pomap.end())
	{
		return it->second;
	}

	GSPixelOffset* off = (GSPixelOffset*)_aligned_malloc(sizeof(GSPixelOffset), 32);

	off->hash = hash;
	off->fbp = fbp;
	off->zbp = zbp;
	off->fpsm = fpsm;
	off->zpsm = zpsm;
	off->bw = bw;

	pixelAddress fpa = m_psm[fpsm].pa;
	pixelAddress zpa = m_psm[zpsm].pa;

	int fs = m_psm[fpsm].bpp >> 5;
	int zs = m_psm[zpsm].bpp >> 5;

	for(int i = 0; i < 2048; i++)
	{
		off->row[i].x = (int)fpa(0, i, fbp, bw) << fs;
		off->row[i].y = (int)zpa(0, i, zbp, bw) << zs;
	}

	for(int i = 0; i < 2048; i++)
	{
		off->col[i].x = m_psm[fpsm].rowOffset[0][i] << fs;
		off->col[i].y = m_psm[zpsm].rowOffset[0][i] << zs;
	}

	m_pomap[hash] = off;

	return off;
}

GSPixelOffset4* GSLocalMemory::GetPixelOffset4(const GIFRegFRAME& FRAME, const GIFRegZBUF& ZBUF)
{
	uint32 fbp = FRAME.Block();
	uint32 zbp = ZBUF.Block();
	uint32 fpsm = FRAME.PSM;
	uint32 zpsm = ZBUF.PSM;
	uint32 bw = FRAME.FBW;

	ASSERT(m_psm[fpsm].trbpp > 8 || m_psm[zpsm].trbpp > 8);

	// "(psm & 0x0f) ^ ((psm & 0xf0) >> 2)" creates 4 bit unique identifiers for render target formats (only)

	uint32 fpsm_hash = (fpsm & 0x0f) ^ ((fpsm & 0x30) >> 2);
	uint32 zpsm_hash = (zpsm & 0x0f) ^ ((zpsm & 0x30) >> 2);

	uint32 hash = (FRAME.FBP << 0) | (ZBUF.ZBP << 9) | (bw << 18) | (fpsm_hash << 24) | (zpsm_hash << 28);

	auto it = m_po4map.find(hash);

	if(it != m_po4map.end())
	{
		return it->second;
	}

	GSPixelOffset4* off = (GSPixelOffset4*)_aligned_malloc(sizeof(GSPixelOffset4), 32);

	off->hash = hash;
	off->fbp = fbp;
	off->zbp = zbp;
	off->fpsm = fpsm;
	off->zpsm = zpsm;
	off->bw = bw;

	pixelAddress fpa = m_psm[fpsm].pa;
	pixelAddress zpa = m_psm[zpsm].pa;

	int fs = m_psm[fpsm].bpp >> 5;
	int zs = m_psm[zpsm].bpp >> 5;

	for(int i = 0; i < 2048; i++)
	{
		off->row[i].x = (int)fpa(0, i, fbp, bw) << fs;
		off->row[i].y = (int)zpa(0, i, zbp, bw) << zs;
	}

	for(int i = 0; i < 512; i++)
	{
		off->col[i].x = m_psm[fpsm].rowOffset[0][i * 4] << fs;
		off->col[i].y = m_psm[zpsm].rowOffset[0][i * 4] << zs;
	}

	m_po4map[hash] = off;

	return off;
}

static bool cmp_vec2x(const GSVector2i& a, const GSVector2i& b) {return a.x < b.x;}

std::vector<GSVector2i>* GSLocalMemory::GetPage2TileMap(const GIFRegTEX0& TEX0)
{
	uint64 hash = TEX0.u64 & 0x3ffffffffull; // TBP0 TBW PSM TW TH

	auto it = m_p2tmap.find(hash);

	if(it != m_p2tmap.end())
	{
		return it->second;
	}

	GSVector2i bs = m_psm[TEX0.PSM].bs;

	int tw = std::max<int>(1 << TEX0.TW, bs.x);
	int th = std::max<int>(1 << TEX0.TH, bs.y);

	const GSOffset* off = GetOffset(TEX0.TBP0, TEX0.TBW, TEX0.PSM);

	std::unordered_map<uint32, std::unordered_set<uint32>> tmp; // key = page, value = y:x, 7 bits each, max 128x128 tiles for the worst case (1024x1024 32bpp 8x8 blocks)

	for(int y = 0; y < th; y += bs.y)
	{
		uint32 base = off->block.row[y >> 3];

		for(int x = 0, i = y << 7; x < tw; x += bs.x, i += bs.x)
		{
			uint32 page = ((base + off->block.col[x >> 3]) >> 5) % MAX_PAGES;

			tmp[page].insert(i >> 3); // ((y << 7) | x) >> 3
		}
	}

	// combine the lower 5 bits of the address into a 9:5 pointer:mask form, so the "valid bits" can be tested against an uint32 array

	auto p2t = new std::vector<GSVector2i>[MAX_PAGES];

	for(const auto &i : tmp)
	{
		uint32 page = i.first;

		auto& tiles = i.second;

		std::unordered_map<uint32, uint32> m;

		for(const auto addr : tiles)
		{
			uint32 row = addr >> 5;
			uint32 col = 1 << (addr & 31);

			auto k = m.find(row);

			if(k != m.end())
			{
				k->second |= col;
			}
			else
			{
				m[row] = col;
			}
		}

		// Allocate vector with initial size
		p2t[page].reserve(m.size());

		// sort by x and flip the mask (it will be used to erase a lot of bits in a loop, [x] &= ~y)

		for(const auto &j : m)
		{
			p2t[page].push_back(GSVector2i(j.first, ~j.second));
		}

		std::sort(p2t[page].begin(), p2t[page].end(), cmp_vec2x);
	}

	m_p2tmap[hash] = p2t;

	return p2t;
}

////////////////////

template<int psm, int bsx, int bsy, int alignment>
void GSLocalMemory::WriteImageColumn(int l, int r, int y, int h, const uint8* src, int srcpitch, const GIFRegBITBLTBUF& BITBLTBUF)
{
	uint32 bp = BITBLTBUF.DBP;
	uint32 bw = BITBLTBUF.DBW;

	const int csy = bsy / 4;

	for(int offset = srcpitch * csy; h >= csy; h -= csy, y += csy, src += offset)
	{
		for(int x = l; x < r; x += bsx)
		{
			switch(psm)
			{
			case PSM_PSMCT32: GSBlock::WriteColumn32<alignment, 0xffffffff>(y, BlockPtr32(x, y, bp, bw), &src[x * 4], srcpitch); break;
			case PSM_PSMCT16: GSBlock::WriteColumn16<alignment>(y, BlockPtr16(x, y, bp, bw), &src[x * 2], srcpitch); break;
			case PSM_PSMCT16S: GSBlock::WriteColumn16<alignment>(y, BlockPtr16S(x, y, bp, bw), &src[x * 2], srcpitch); break;
			case PSM_PSMT8: GSBlock::WriteColumn8<alignment>(y, BlockPtr8(x, y, bp, bw), &src[x], srcpitch); break;
			case PSM_PSMT4: GSBlock::WriteColumn4<alignment>(y, BlockPtr4(x, y, bp, bw), &src[x >> 1], srcpitch); break;
			case PSM_PSMZ32: GSBlock::WriteColumn32<alignment, 0xffffffff>(y, BlockPtr32Z(x, y, bp, bw), &src[x * 4], srcpitch); break;
			case PSM_PSMZ16: GSBlock::WriteColumn16<alignment>(y, BlockPtr16Z(x, y, bp, bw), &src[x * 2], srcpitch); break;
			case PSM_PSMZ16S: GSBlock::WriteColumn16<alignment>(y, BlockPtr16SZ(x, y, bp, bw), &src[x * 2], srcpitch); break;
			// TODO
			default: __assume(0);
			}
		}
	}
}

template<int psm, int bsx, int bsy, int alignment>
void GSLocalMemory::WriteImageBlock(int l, int r, int y, int h, const uint8* src, int srcpitch, const GIFRegBITBLTBUF& BITBLTBUF)
{
	uint32 bp = BITBLTBUF.DBP;
	uint32 bw = BITBLTBUF.DBW;

	for(int offset = srcpitch * bsy; h >= bsy; h -= bsy, y += bsy, src += offset)
	{
		for(int x = l; x < r; x += bsx)
		{
			switch(psm)
			{
			case PSM_PSMCT32: GSBlock::WriteBlock32<alignment, 0xffffffff>(BlockPtr32(x, y, bp, bw), &src[x * 4], srcpitch); break;
			case PSM_PSMCT16: GSBlock::WriteBlock16<alignment>(BlockPtr16(x, y, bp, bw), &src[x * 2], srcpitch); break;
			case PSM_PSMCT16S: GSBlock::WriteBlock16<alignment>(BlockPtr16S(x, y, bp, bw), &src[x * 2], srcpitch); break;
			case PSM_PSMT8: GSBlock::WriteBlock8<alignment>(BlockPtr8(x, y, bp, bw), &src[x], srcpitch); break;
			case PSM_PSMT4: GSBlock::WriteBlock4<alignment>(BlockPtr4(x, y, bp, bw), &src[x >> 1], srcpitch); break;
			case PSM_PSMZ32: GSBlock::WriteBlock32<alignment, 0xffffffff>(BlockPtr32Z(x, y, bp, bw), &src[x * 4], srcpitch); break;
			case PSM_PSMZ16: GSBlock::WriteBlock16<alignment>(BlockPtr16Z(x, y, bp, bw), &src[x * 2], srcpitch); break;
			case PSM_PSMZ16S: GSBlock::WriteBlock16<alignment>(BlockPtr16SZ(x, y, bp, bw), &src[x * 2], srcpitch); break;
			// TODO
			default: __assume(0);
			}
		}
	}
}

template<int psm, int bsx, int bsy>
void GSLocalMemory::WriteImageLeftRight(int l, int r, int y, int h, const uint8* src, int srcpitch, const GIFRegBITBLTBUF& BITBLTBUF)
{
	uint32 bp = BITBLTBUF.DBP;
	uint32 bw = BITBLTBUF.DBW;

	for(; h > 0; y++, h--, src += srcpitch)
	{
		for(int x = l; x < r; x++)
		{
			switch(psm)
			{
			case PSM_PSMCT32: WritePixel32(x, y, *(uint32*)&src[x * 4], bp, bw); break;
			case PSM_PSMCT16: WritePixel16(x, y, *(uint16*)&src[x * 2], bp, bw); break;
			case PSM_PSMCT16S: WritePixel16S(x, y, *(uint16*)&src[x * 2], bp, bw); break;
			case PSM_PSMT8: WritePixel8(x, y, src[x], bp, bw); break;
			case PSM_PSMT4: WritePixel4(x, y, src[x >> 1] >> ((x & 1) << 2), bp, bw); break;
			case PSM_PSMZ32: WritePixel32Z(x, y, *(uint32*)&src[x * 4], bp, bw); break;
			case PSM_PSMZ16: WritePixel16Z(x, y, *(uint16*)&src[x * 2], bp, bw); break;
			case PSM_PSMZ16S: WritePixel16SZ(x, y, *(uint16*)&src[x * 2], bp, bw); break;
			// TODO
			default: __assume(0);
			}
		}
	}
}

template<int psm, int bsx, int bsy, int trbpp>
void GSLocalMemory::WriteImageTopBottom(int l, int r, int y, int h, const uint8* src, int srcpitch, const GIFRegBITBLTBUF& BITBLTBUF)
{
	alignas(32) uint8 buff[64]; // merge buffer for one column

	uint32 bp = BITBLTBUF.DBP;
	uint32 bw = BITBLTBUF.DBW;

	const int csy = bsy / 4;

	// merge incomplete column

	int y2 = y & (csy - 1);

	if(y2 > 0)
	{
		int h2 = std::min(h, csy - y2);

		for(int x = l; x < r; x += bsx)
		{
			uint8* dst = NULL;

			switch(psm)
			{
			case PSM_PSMCT32: dst = BlockPtr32(x, y, bp, bw); break;
			case PSM_PSMCT16: dst = BlockPtr16(x, y, bp, bw); break;
			case PSM_PSMCT16S: dst = BlockPtr16S(x, y, bp, bw); break;
			case PSM_PSMT8: dst = BlockPtr8(x, y, bp, bw); break;
			case PSM_PSMT4: dst = BlockPtr4(x, y, bp, bw); break;
			case PSM_PSMZ32: dst = BlockPtr32Z(x, y, bp, bw); break;
			case PSM_PSMZ16: dst = BlockPtr16Z(x, y, bp, bw); break;
			case PSM_PSMZ16S: dst = BlockPtr16SZ(x, y, bp, bw); break;
			// TODO
			default: __assume(0);
			}

			switch(psm)
			{
			case PSM_PSMCT32:
			case PSM_PSMZ32:
				GSBlock::ReadColumn32(y, dst, buff, 32);
				memcpy(&buff[32], &src[x * 4], 32);
				GSBlock::WriteColumn32<32, 0xffffffff>(y, dst, buff, 32);
				break;
			case PSM_PSMCT16:
			case PSM_PSMCT16S:
			case PSM_PSMZ16:
			case PSM_PSMZ16S:
				GSBlock::ReadColumn16(y, dst, buff, 32);
				memcpy(&buff[32], &src[x * 2], 32);
				GSBlock::WriteColumn16<32>(y, dst, buff, 32);
				break;
			case PSM_PSMT8:
				GSBlock::ReadColumn8(y, dst, buff, 16);
				for(int i = 0, j = y2; i < h2; i++, j++) memcpy(&buff[j * 16], &src[i * srcpitch + x], 16);
				GSBlock::WriteColumn8<32>(y, dst, buff, 16);
				break;
			case PSM_PSMT4:
				GSBlock::ReadColumn4(y, dst, buff, 16);
				for(int i = 0, j = y2; i < h2; i++, j++) memcpy(&buff[j * 16], &src[i * srcpitch + (x >> 1)], 16);
				GSBlock::WriteColumn4<32>(y, dst, buff, 16);
				break;
			// TODO
			default:
				__assume(0);
			}
		}

		src += srcpitch * h2;
		y += h2;
		h -= h2;
	}

	// write whole columns

	{
		int h2 = h & ~(csy - 1);

		if(h2 > 0)
		{
			size_t addr = (size_t)&src[l * trbpp >> 3];

			if((addr & 31) == 0 && (srcpitch & 31) == 0)
			{
				WriteImageColumn<psm, bsx, bsy, 32>(l, r, y, h2, src, srcpitch, BITBLTBUF);
			}
			else if((addr & 15) == 0 && (srcpitch & 15) == 0)
			{
				WriteImageColumn<psm, bsx, bsy, 16>(l, r, y, h2, src, srcpitch, BITBLTBUF);
			}
			else
			{
				WriteImageColumn<psm, bsx, bsy, 0>(l, r, y, h2, src, srcpitch, BITBLTBUF);
			}

			src += srcpitch * h2;
			y += h2;
			h -= h2;
		}
	}

	// merge incomplete column

	if(h >= 1)
	{
		for(int x = l; x < r; x += bsx)
		{
			uint8* dst = NULL;

			switch(psm)
			{
			case PSM_PSMCT32: dst = BlockPtr32(x, y, bp, bw); break;
			case PSM_PSMCT16: dst = BlockPtr16(x, y, bp, bw); break;
			case PSM_PSMCT16S: dst = BlockPtr16S(x, y, bp, bw); break;
			case PSM_PSMT8: dst = BlockPtr8(x, y, bp, bw); break;
			case PSM_PSMT4: dst = BlockPtr4(x, y, bp, bw); break;
			case PSM_PSMZ32: dst = BlockPtr32Z(x, y, bp, bw); break;
			case PSM_PSMZ16: dst = BlockPtr16Z(x, y, bp, bw); break;
			case PSM_PSMZ16S: dst = BlockPtr16SZ(x, y, bp, bw); break;
			// TODO
			default: __assume(0);
			}

			switch(psm)
			{
			case PSM_PSMCT32:
			case PSM_PSMZ32:
				GSBlock::ReadColumn32(y, dst, buff, 32);
				memcpy(&buff[0], &src[x * 4], 32);
				GSBlock::WriteColumn32<32, 0xffffffff>(y, dst, buff, 32);
				break;
			case PSM_PSMCT16:
			case PSM_PSMCT16S:
			case PSM_PSMZ16:
			case PSM_PSMZ16S:
				GSBlock::ReadColumn16(y, dst, buff, 32);
				memcpy(&buff[0], &src[x * 2], 32);
				GSBlock::WriteColumn16<32>(y, dst, buff, 32);
				break;
			case PSM_PSMT8:
				GSBlock::ReadColumn8(y, dst, buff, 16);
				for(int i = 0; i < h; i++) memcpy(&buff[i * 16], &src[i * srcpitch + x], 16);
				GSBlock::WriteColumn8<32>(y, dst, buff, 16);
				break;
			case PSM_PSMT4:
				GSBlock::ReadColumn4(y, dst, buff, 16);
				for(int i = 0; i < h; i++) memcpy(&buff[i * 16], &src[i * srcpitch + (x >> 1)], 16);
				GSBlock::WriteColumn4<32>(y, dst, buff, 16);
				break;
			// TODO
			default:
				__assume(0);
			}
		}
	}
}

template<int psm, int bsx, int bsy, int trbpp>
void GSLocalMemory::WriteImage(int& tx, int& ty, const uint8* src, int len, GIFRegBITBLTBUF& BITBLTBUF, GIFRegTRXPOS& TRXPOS, GIFRegTRXREG& TRXREG)
{
	if(TRXREG.RRW == 0) return;

	int l = (int)TRXPOS.DSAX;
	int r = l + (int)TRXREG.RRW;

	// finish the incomplete row first

	if(tx != l)
	{
		int n = std::min(len, (r - tx) * trbpp >> 3);
		WriteImageX(tx, ty, src, n, BITBLTBUF, TRXPOS, TRXREG);
		src += n;
		len -= n;
	}

	int la = (l + (bsx - 1)) & ~(bsx - 1);
	int ra = r & ~(bsx - 1);
	int srcpitch = (r - l) * trbpp >> 3;
	int h = len / srcpitch;

	if(ra - la >= bsx && h > 0) // "transfer width" >= "block width" && there is at least one full row
	{
		const uint8* s = &src[-l * trbpp >> 3];

		src += srcpitch * h;
		len -= srcpitch * h;

		// left part

		if(l < la)
		{
			WriteImageLeftRight<psm, bsx, bsy>(l, la, ty, h, s, srcpitch, BITBLTBUF);
		}

		// right part

		if(ra < r)
		{
			WriteImageLeftRight<psm, bsx, bsy>(ra, r, ty, h, s, srcpitch, BITBLTBUF);
		}

		// horizontally aligned part

		if(la < ra)
		{
			// top part

			{
				int h2 = std::min(h, bsy - (ty & (bsy - 1)));

				if(h2 < bsy)
				{
					WriteImageTopBottom<psm, bsx, bsy, trbpp>(la, ra, ty, h2, s, srcpitch, BITBLTBUF);

					s += srcpitch * h2;
					ty += h2;
					h -= h2;
				}
			}

			// horizontally and vertically aligned part

			{
				int h2 = h & ~(bsy - 1);

				if(h2 > 0)
				{
					size_t addr = (size_t)&s[la * trbpp >> 3];

					if((addr & 31) == 0 && (srcpitch & 31) == 0)
					{
						WriteImageBlock<psm, bsx, bsy, 32>(la, ra, ty, h2, s, srcpitch, BITBLTBUF);
					}
					else if((addr & 15) == 0 && (srcpitch & 15) == 0)
					{
						WriteImageBlock<psm, bsx, bsy, 16>(la, ra, ty, h2, s, srcpitch, BITBLTBUF);
					}
					else
					{
						WriteImageBlock<psm, bsx, bsy, 0>(la, ra, ty, h2, s, srcpitch, BITBLTBUF);
					}

					s += srcpitch * h2;
					ty += h2;
					h -= h2;
				}
			}

			// bottom part

			if(h > 0)
			{
				WriteImageTopBottom<psm, bsx, bsy, trbpp>(la, ra, ty, h, s, srcpitch, BITBLTBUF);

				// s += srcpitch * h;
				ty += h;
				// h -= h;
			}
		}
	}

	// the rest

	if(len > 0)
	{
		WriteImageX(tx, ty, src, len, BITBLTBUF, TRXPOS, TRXREG);
	}
}


static bool IsTopLeftAligned(int dsax, int tx, int ty, int bw, int bh)
{
	return ((dsax & (bw-1)) == 0 && (tx & (bw-1)) == 0 && dsax == tx && (ty & (bh-1)) == 0);
}

void GSLocalMemory::WriteImage24(int& tx, int& ty, const uint8* src, int len, GIFRegBITBLTBUF& BITBLTBUF, GIFRegTRXPOS& TRXPOS, GIFRegTRXREG& TRXREG)
{
	if(TRXREG.RRW == 0) return;

	uint32 bp = BITBLTBUF.DBP;
	uint32 bw = BITBLTBUF.DBW;

	int tw = TRXPOS.DSAX + TRXREG.RRW, srcpitch = TRXREG.RRW * 3;
	int th = len / srcpitch;

	bool aligned = IsTopLeftAligned(TRXPOS.DSAX, tx, ty, 8, 8);

	if(!aligned || (tw & 7) || (th & 7) || (len % srcpitch))
	{
		// TODO

		WriteImageX(tx, ty, src, len, BITBLTBUF, TRXPOS, TRXREG);
	}
	else
	{
		th += ty;

		for(int y = ty; y < th; y += 8, src += srcpitch * 8)
		{
			for(int x = tx; x < tw; x += 8)
			{
				GSBlock::UnpackAndWriteBlock24(src + (x - tx) * 3, srcpitch, BlockPtr32(x, y, bp, bw));
			}
		}

		ty = th;
	}
}

void GSLocalMemory::WriteImage8H(int& tx, int& ty, const uint8* src, int len, GIFRegBITBLTBUF& BITBLTBUF, GIFRegTRXPOS& TRXPOS, GIFRegTRXREG& TRXREG)
{
	if(TRXREG.RRW == 0) return;

	uint32 bp = BITBLTBUF.DBP;
	uint32 bw = BITBLTBUF.DBW;

	int tw = TRXPOS.DSAX + TRXREG.RRW, srcpitch = TRXREG.RRW;
	int th = len / srcpitch;

	bool aligned = IsTopLeftAligned(TRXPOS.DSAX, tx, ty, 8, 8);

	if(!aligned || (tw & 7) || (th & 7) || (len % srcpitch))
	{
		// TODO

		WriteImageX(tx, ty, src, len, BITBLTBUF, TRXPOS, TRXREG);
	}
	else
	{
		th += ty;

		for(int y = ty; y < th; y += 8, src += srcpitch * 8)
		{
			for(int x = tx; x < tw; x += 8)
			{
				GSBlock::UnpackAndWriteBlock8H(src + (x - tx), srcpitch, BlockPtr32(x, y, bp, bw));
			}
		}

		ty = th;
	}
}

void GSLocalMemory::WriteImage4HL(int& tx, int& ty, const uint8* src, int len, GIFRegBITBLTBUF& BITBLTBUF, GIFRegTRXPOS& TRXPOS, GIFRegTRXREG& TRXREG)
{
	if(TRXREG.RRW == 0) return;

	uint32 bp = BITBLTBUF.DBP;
	uint32 bw = BITBLTBUF.DBW;

	int tw = TRXPOS.DSAX + TRXREG.RRW, srcpitch = TRXREG.RRW / 2;
	int th = len / srcpitch;

	bool aligned = IsTopLeftAligned(TRXPOS.DSAX, tx, ty, 8, 8);

	if(!aligned || (tw & 7) || (th & 7) || (len % srcpitch))
	{
		// TODO

		WriteImageX(tx, ty, src, len, BITBLTBUF, TRXPOS, TRXREG);
	}
	else
	{
		th += ty;

		for(int y = ty; y < th; y += 8, src += srcpitch * 8)
		{
			for(int x = tx; x < tw; x += 8)
			{
				GSBlock::UnpackAndWriteBlock4HL(src + (x - tx) / 2, srcpitch, BlockPtr32(x, y, bp, bw));
			}
		}

		ty = th;
	}
}

void GSLocalMemory::WriteImage4HH(int& tx, int& ty, const uint8* src, int len, GIFRegBITBLTBUF& BITBLTBUF, GIFRegTRXPOS& TRXPOS, GIFRegTRXREG& TRXREG)
{
	if(TRXREG.RRW == 0) return;

	uint32 bp = BITBLTBUF.DBP;
	uint32 bw = BITBLTBUF.DBW;

	int tw = TRXPOS.DSAX + TRXREG.RRW, srcpitch = TRXREG.RRW / 2;
	int th = len / srcpitch;

	bool aligned = IsTopLeftAligned(TRXPOS.DSAX, tx, ty, 8, 8);

	if(!aligned || (tw & 7) || (th & 7) || (len % srcpitch))
	{
		// TODO

		WriteImageX(tx, ty, src, len, BITBLTBUF, TRXPOS, TRXREG);
	}
	else
	{
		th += ty;

		for(int y = ty; y < th; y += 8, src += srcpitch * 8)
		{
			for(int x = tx; x < tw; x += 8)
			{
				GSBlock::UnpackAndWriteBlock4HH(src + (x - tx) / 2, srcpitch, BlockPtr32(x, y, bp, bw));
			}
		}

		ty = th;
	}
}

void GSLocalMemory::WriteImage24Z(int& tx, int& ty, const uint8* src, int len, GIFRegBITBLTBUF& BITBLTBUF, GIFRegTRXPOS& TRXPOS, GIFRegTRXREG& TRXREG)
{
	if(TRXREG.RRW == 0) return;

	uint32 bp = BITBLTBUF.DBP;
	uint32 bw = BITBLTBUF.DBW;

	int tw = TRXPOS.DSAX + TRXREG.RRW, srcpitch = TRXREG.RRW * 3;
	int th = len / srcpitch;

	bool aligned = IsTopLeftAligned(TRXPOS.DSAX, tx, ty, 8, 8);

	if(!aligned || (tw & 7) || (th & 7) || (len % srcpitch))
	{
		// TODO

		WriteImageX(tx, ty, src, len, BITBLTBUF, TRXPOS, TRXREG);
	}
	else
	{
		th += ty;

		for(int y = ty; y < th; y += 8, src += srcpitch * 8)
		{
			for(int x = tx; x < tw; x += 8)
			{
				GSBlock::UnpackAndWriteBlock24(src + (x - tx) * 3, srcpitch, BlockPtr32Z(x, y, bp, bw));
			}
		}

		ty = th;
	}
}

void GSLocalMemory::WriteImageX(int& tx, int& ty, const uint8* src, int len, GIFRegBITBLTBUF& BITBLTBUF, GIFRegTRXPOS& TRXPOS, GIFRegTRXREG& TRXREG)
{
	if(len <= 0) return;

	const uint8* pb = (uint8*)src;
	const uint16* pw = (uint16*)src;
	const uint32* pd = (uint32*)src;

	uint32 bp = BITBLTBUF.DBP;
	uint32 bw = BITBLTBUF.DBW;
	psm_t* psm = &m_psm[BITBLTBUF.DPSM];

	int x = tx;
	int y = ty;
	int sx = (int)TRXPOS.DSAX;
	int ex = sx + (int)TRXREG.RRW;

	switch(BITBLTBUF.DPSM)
	{
	case PSM_PSMCT32:
	case PSM_PSMZ32:

		len /= 4;

		while(len > 0)
		{
			uint32 addr = psm->pa(0, y, bp, bw);
			int* offset = psm->rowOffset[y & 7];

			for(; len > 0 && x < ex; len--, x++, pd++)
			{
				WritePixel32(addr + offset[x], *pd);
			}

			if(x >= ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMCT24:
	case PSM_PSMZ24:

		len /= 3;

		while(len > 0)
		{
			uint32 addr = psm->pa(0, y, bp, bw);
			int* offset = psm->rowOffset[y & 7];

			for(; len > 0 && x < ex; len--, x++, pb += 3)
			{
				WritePixel24(addr + offset[x], *(uint32*)pb);
			}

			if(x >= ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMCT16:
	case PSM_PSMCT16S:
	case PSM_PSMZ16:
	case PSM_PSMZ16S:

		len /= 2;

		while(len > 0)
		{
			uint32 addr = psm->pa(0, y, bp, bw);
			int* offset = psm->rowOffset[y & 7];

			for(; len > 0 && x < ex; len--, x++, pw++)
			{
				WritePixel16(addr + offset[x], *pw);
			}

			if(x >= ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMT8:

		while(len > 0)
		{
			uint32 addr = psm->pa(0, y, bp, bw);
			int* offset = psm->rowOffset[y & 7];

			for(; len > 0 && x < ex; len--, x++, pb++)
			{
				WritePixel8(addr + offset[x], *pb);
			}

			if(x >= ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMT4:

		while(len > 0)
		{
			uint32 addr = psm->pa(0, y, bp, bw);
			int* offset = psm->rowOffset[y & 7];

			for(; len > 0 && x < ex; len--, x += 2, pb++)
			{
				WritePixel4(addr + offset[x + 0], *pb & 0xf);
				WritePixel4(addr + offset[x + 1], *pb >> 4);
			}

			if(x >= ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMT8H:

		while(len > 0)
		{
			uint32 addr = psm->pa(0, y, bp, bw);
			int* offset = psm->rowOffset[y & 7];

			for(; len > 0 && x < ex; len--, x++, pb++)
			{
				WritePixel8H(addr + offset[x], *pb);
			}

			if(x >= ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMT4HL:

		while(len > 0)
		{
			uint32 addr = psm->pa(0, y, bp, bw);
			int* offset = psm->rowOffset[y & 7];

			for(; len > 0 && x < ex; len--, x += 2, pb++)
			{
				WritePixel4HL(addr + offset[x + 0], *pb & 0xf);
				WritePixel4HL(addr + offset[x + 1], *pb >> 4);
			}

			if(x >= ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMT4HH:

		while(len > 0)
		{
			uint32 addr = psm->pa(0, y, bp, bw);
			int* offset = psm->rowOffset[y & 7];

			for(; len > 0 && x < ex; len--, x += 2, pb++)
			{
				WritePixel4HH(addr + offset[x + 0], *pb & 0xf);
				WritePixel4HH(addr + offset[x + 1], *pb >> 4);
			}

			if(x >= ex) {x = sx; y++;}
		}

		break;
	}

	tx = x;
	ty = y;
}

//

void GSLocalMemory::ReadImageX(int& tx, int& ty, uint8* dst, int len, GIFRegBITBLTBUF& BITBLTBUF, GIFRegTRXPOS& TRXPOS, GIFRegTRXREG& TRXREG) const
{
	if(len <= 0) return;

	uint8* RESTRICT pb = (uint8*)dst;
	uint16* RESTRICT pw = (uint16*)dst;
	uint32* RESTRICT pd = (uint32*)dst;

	uint32 bp = BITBLTBUF.SBP;
	uint32 bw = BITBLTBUF.SBW;
	psm_t* RESTRICT psm = &m_psm[BITBLTBUF.SPSM];

	int x = tx;
	int y = ty;
	int sx = (int)TRXPOS.SSAX;
	int ex = sx + (int)TRXREG.RRW;

	// printf("spsm=%d x=%d ex=%d y=%d len=%d\n", BITBLTBUF.SPSM, x, ex, y, len);

	switch(BITBLTBUF.SPSM)
	{
	case PSM_PSMCT32:
	case PSM_PSMZ32:

		// MGS1 intro, fade effect between two scenes (airplane outside-inside transition)

		len /= 4;

		while(len > 0)
		{
			int* RESTRICT offset = psm->rowOffset[y & 7];
			uint32* RESTRICT ps = &m_vm32[psm->pa(0, y, bp, bw)];

			for(; len > 0 && x < ex && (x & 7); len--, x++, pd++) 
			{
				*pd = ps[offset[x]];
			}

			// aligned to a column

			for(int ex8 = ex - 8; len >= 8 && x <= ex8; len -= 8, x += 8, pd += 8)
			{
				int off = offset[x];

				GSVector4i::store<false>(&pd[0], GSVector4i::load(&ps[off + 0], &ps[off + 4]));
				GSVector4i::store<false>(&pd[4], GSVector4i::load(&ps[off + 8], &ps[off + 12]));

				for(int i = 0; i < 8; i++) ASSERT(pd[i] == ps[offset[x + i]]);
			}

			for(; len > 0 && x < ex; len--, x++, pd++)
			{
				*pd = ps[offset[x]];
			}

			if(x == ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMCT24:
	case PSM_PSMZ24:

		len /= 3;

		while(len > 0)
		{
			int* RESTRICT offset = psm->rowOffset[y & 7];
			uint32* RESTRICT ps = &m_vm32[psm->pa(0, y, bp, bw)];

			for(; len > 0 && x < ex; len--, x++, pb += 3)
			{
				uint32 c = ps[offset[x]];

				pb[0] = (uint8)(c);
				pb[1] = (uint8)(c >> 8);
				pb[2] = (uint8)(c >> 16);
			}

			if(x == ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMCT16:
	case PSM_PSMCT16S:
	case PSM_PSMZ16:
	case PSM_PSMZ16S:

		len /= 2;

		while(len > 0)
		{
			int* RESTRICT offset = psm->rowOffset[y & 7];
			uint16* RESTRICT ps = &m_vm16[psm->pa(0, y, bp, bw)];

			for(int ex4 = ex - 4; len >= 4 && x <= ex4; len -= 4, x += 4, pw += 4)
			{
				pw[0] = ps[offset[x + 0]];
				pw[1] = ps[offset[x + 1]];
				pw[2] = ps[offset[x + 2]];
				pw[3] = ps[offset[x + 3]];
			}

			for(; len > 0 && x < ex; len--, x++, pw++)
			{
				*pw = ps[offset[x]];
			}

			if(x == ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMT8:

		while(len > 0)
		{
			int* RESTRICT offset = psm->rowOffset[y & 7];
			uint8* RESTRICT ps = &m_vm8[psm->pa(0, y, bp, bw)];

			for(int ex4 = ex - 4; len >= 4 && x <= ex4; len -= 4, x += 4, pb += 4)
			{
				pb[0] = ps[offset[x + 0]];
				pb[1] = ps[offset[x + 1]];
				pb[2] = ps[offset[x + 2]];
				pb[3] = ps[offset[x + 3]];
			}

			for(; len > 0 && x < ex; len--, x++, pb++)
			{
				*pb = ps[offset[x]];
			}

			if(x == ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMT4:

		while(len > 0)
		{
			uint32 addr = psm->pa(0, y, bp, bw);
			int* RESTRICT offset = psm->rowOffset[y & 7];

			for(; len > 0 && x < ex; len--, x += 2, pb++)
			{
				*pb = (uint8)(ReadPixel4(addr + offset[x + 0]) | (ReadPixel4(addr + offset[x + 1]) << 4));
			}

			if(x == ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMT8H:

		while(len > 0)
		{
			int* RESTRICT offset = psm->rowOffset[y & 7];
			uint32* RESTRICT ps = &m_vm32[psm->pa(0, y, bp, bw)];

			for(int ex4 = ex - 4; len >= 4 && x <= ex4; len -= 4, x += 4, pb += 4)
			{
				pb[0] = (uint8)(ps[offset[x + 0]] >> 24);
				pb[1] = (uint8)(ps[offset[x + 1]] >> 24);
				pb[2] = (uint8)(ps[offset[x + 2]] >> 24);
				pb[3] = (uint8)(ps[offset[x + 3]] >> 24);
			}

			for(; len > 0 && x < ex; len--, x++, pb++)
			{
				*pb = (uint8)(ps[offset[x]] >> 24);
			}

			if(x == ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMT4HL:

		while(len > 0)
		{
			int* offset = psm->rowOffset[y & 7];
			uint32* RESTRICT ps = &m_vm32[psm->pa(0, y, bp, bw)];

			for(; len > 0 && x < ex; len--, x += 2, pb++)
			{
				uint32 c0 = (ps[offset[x + 0]] >> 24) & 0x0f;
				uint32 c1 = (ps[offset[x + 1]] >> 20) & 0xf0;

				*pb = (uint8)(c0 | c1);
			}

			if(x == ex) {x = sx; y++;}
		}

		break;

	case PSM_PSMT4HH:

		while(len > 0)
		{
			int* RESTRICT offset = psm->rowOffset[y & 7];
			uint32* RESTRICT ps = &m_vm32[psm->pa(0, y, bp, bw)];

			for(; len > 0 && x < ex; len--, x += 2, pb++)
			{
				uint32 c0 = (ps[offset[x + 0]] >> 28) & 0x0f;
				uint32 c1 = (ps[offset[x + 1]] >> 24) & 0xf0;

				*pb = (uint8)(c0 | c1);
			}

			if(x == ex) {x = sx; y++;}
		}

		break;
	}

	tx = x;
	ty = y;
}

///////////////////

void GSLocalMemory::ReadTexture32(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	FOREACH_BLOCK_START(r, 8, 8, 32)
	{
		GSBlock::ReadBlock32(src, read_dst, dstpitch);
	}
	FOREACH_BLOCK_END
}

void GSLocalMemory::ReadTexture24(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	if(TEXA.AEM)
	{
		FOREACH_BLOCK_START(r, 8, 8, 32)
		{
			GSBlock::ReadAndExpandBlock24<true>(src, read_dst, dstpitch, TEXA);
		}
		FOREACH_BLOCK_END
	}
	else
	{
		FOREACH_BLOCK_START(r, 8, 8, 32)
		{
			GSBlock::ReadAndExpandBlock24<false>(src, read_dst, dstpitch, TEXA);
		}
		FOREACH_BLOCK_END
	}
}

void GSLocalMemory::ReadTextureGPU24(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	FOREACH_BLOCK_START(r, 16, 8, 16)
	{
		GSBlock::ReadBlock16(src, read_dst, dstpitch);
	}
	FOREACH_BLOCK_END

	// Convert packed RGB scanline to 32 bits RGBA
	ASSERT(dstpitch >= r.width() * 4);
	for(int y = r.top; y < r.bottom; y ++) {
		uint8* line = dst + y * dstpitch;

		for(int x = r.right; x >= r.left; x--) {
			*(uint32*)&line[x * 4] = *(uint32*)&line[x * 3] & 0xFFFFFF;
		}
	}
}

void GSLocalMemory::ReadTexture16(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	if(TEXA.AEM)
	{
		FOREACH_BLOCK_START(r, 16, 8, 32)
		{
			GSBlock::ReadAndExpandBlock16<true>(src, read_dst, dstpitch, TEXA);
		}
		FOREACH_BLOCK_END
	}
	else
	{
		FOREACH_BLOCK_START(r, 16, 8, 32)
		{
			GSBlock::ReadAndExpandBlock16<false>(src, read_dst, dstpitch, TEXA);
		}
		FOREACH_BLOCK_END
	}
}

void GSLocalMemory::ReadTexture8(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	const uint32* pal = m_clut;

	FOREACH_BLOCK_START(r, 16, 16, 32)
	{
		GSBlock::ReadAndExpandBlock8_32(src, read_dst, dstpitch, pal);
	}
	FOREACH_BLOCK_END
}

void GSLocalMemory::ReadTexture4(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	const uint64* pal = m_clut;

	FOREACH_BLOCK_START(r, 32, 16, 32)
	{
		GSBlock::ReadAndExpandBlock4_32(src, read_dst, dstpitch, pal);
	}
	FOREACH_BLOCK_END
}

void GSLocalMemory::ReadTexture8H(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	const uint32* pal = m_clut;

	FOREACH_BLOCK_START(r, 8, 8, 32)
	{
		GSBlock::ReadAndExpandBlock8H_32(src, read_dst, dstpitch, pal);
	}
	FOREACH_BLOCK_END
}

void GSLocalMemory::ReadTexture4HL(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	const uint32* pal = m_clut;

	FOREACH_BLOCK_START(r, 8, 8, 32)
	{
		GSBlock::ReadAndExpandBlock4HL_32(src, read_dst, dstpitch, pal);
	}
	FOREACH_BLOCK_END
}

void GSLocalMemory::ReadTexture4HH(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	const uint32* pal = m_clut;

	FOREACH_BLOCK_START(r, 8, 8, 32)
	{
		GSBlock::ReadAndExpandBlock4HH_32(src, read_dst, dstpitch, pal);
	}
	FOREACH_BLOCK_END
}

///////////////////

void GSLocalMemory::ReadTextureBlock32(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	GSBlock::ReadBlock32(BlockPtr(bp), dst, dstpitch);
}

void GSLocalMemory::ReadTextureBlock24(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	if(TEXA.AEM)
	{
		GSBlock::ReadAndExpandBlock24<true>(BlockPtr(bp), dst, dstpitch, TEXA);
	}
	else
	{
		GSBlock::ReadAndExpandBlock24<false>(BlockPtr(bp), dst, dstpitch, TEXA);
	}
}

void GSLocalMemory::ReadTextureBlock16(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	if(TEXA.AEM)
	{
		GSBlock::ReadAndExpandBlock16<true>(BlockPtr(bp), dst, dstpitch, TEXA);
	}
	else
	{
		GSBlock::ReadAndExpandBlock16<false>(BlockPtr(bp), dst, dstpitch, TEXA);
	}
}

void GSLocalMemory::ReadTextureBlock8(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	GSBlock::ReadAndExpandBlock8_32(BlockPtr(bp), dst, dstpitch, m_clut);
}

void GSLocalMemory::ReadTextureBlock4(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	GSBlock::ReadAndExpandBlock4_32(BlockPtr(bp), dst, dstpitch, m_clut);
}

void GSLocalMemory::ReadTextureBlock8H(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	GSBlock::ReadAndExpandBlock8H_32(BlockPtr(bp), dst, dstpitch, m_clut);
}

void GSLocalMemory::ReadTextureBlock4HL(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	GSBlock::ReadAndExpandBlock4HL_32(BlockPtr(bp), dst, dstpitch, m_clut);
}

void GSLocalMemory::ReadTextureBlock4HH(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	GSBlock::ReadAndExpandBlock4HH_32(BlockPtr(bp), dst, dstpitch, m_clut);
}

///////////////////

void GSLocalMemory::ReadTexture(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	const psm_t& psm = m_psm[off->psm];

	readTexel rt = psm.rt;
	readTexture rtx = psm.rtx;

	if(r.width() < psm.bs.x || r.height() < psm.bs.y
	|| (r.left & (psm.bs.x - 1)) || (r.top & (psm.bs.y - 1))
	|| (r.right & (psm.bs.x - 1)) || (r.bottom & (psm.bs.y - 1)))
	{
		GIFRegTEX0 TEX0;

		TEX0.TBP0 = off->bp;
		TEX0.TBW = off->bw;
		TEX0.PSM = off->psm;

		GSVector4i cr = r.ralign<Align_Inside>(psm.bs);

		bool aligned = ((size_t)(dst + (cr.left - r.left) * sizeof(uint32)) & 0xf) == 0;

		if(cr.rempty() || !aligned)
		{
			// TODO: expand r to block size, read into temp buffer

			if(!aligned) printf("unaligned memory pointer passed to ReadTexture\n");

			for(int y = r.top; y < r.bottom; y++, dst += dstpitch)
			{
				for(int x = r.left, i = 0; x < r.right; x++, i++)
				{
					((uint32*)dst)[i] = (this->*rt)(x, y, TEX0, TEXA);
				}
			}
		}
		else
		{
			for(int y = r.top; y < cr.top; y++, dst += dstpitch)
			{
				for(int x = r.left, i = 0; x < r.right; x++, i++)
				{
					((uint32*)dst)[i] = (this->*rt)(x, y, TEX0, TEXA);
				}
			}

			for(int y = cr.bottom; y < r.bottom; y++, dst += dstpitch)
			{
				for(int x = r.left, i = 0; x < r.right; x++, i++)
				{
					((uint32*)dst)[i] = (this->*rt)(x, y, TEX0, TEXA);
				}
			}

			for(int y = cr.top; y < cr.bottom; y++, dst += dstpitch)
			{
				for(int x = r.left, i = 0; x < cr.left; x++, i++)
				{
					((uint32*)dst)[i] = (this->*rt)(x, y, TEX0, TEXA);
				}

				for(int x = cr.right, i = x - r.left; x < r.right; x++, i++)
				{
					((uint32*)dst)[i] = (this->*rt)(x, y, TEX0, TEXA);
				}
			}

			if(!cr.rempty())
			{
				(this->*rtx)(off, cr, dst + (cr.left - r.left) * sizeof(uint32), dstpitch, TEXA);
			}
		}
	}
	else
	{
		(this->*rtx)(off, r, dst, dstpitch, TEXA);
	}
}

// 32/8

void GSLocalMemory::ReadTexture8P(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	FOREACH_BLOCK_START(r, 16, 16, 8)
	{
		GSBlock::ReadBlock8(src, read_dst, dstpitch);
	}
	FOREACH_BLOCK_END
}

void GSLocalMemory::ReadTexture4P(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	FOREACH_BLOCK_START(r, 32, 16, 8)
	{
		GSBlock::ReadBlock4P(src, read_dst, dstpitch);
	}
	FOREACH_BLOCK_END
}

void GSLocalMemory::ReadTexture8HP(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	FOREACH_BLOCK_START(r, 8, 8, 8)
	{
		GSBlock::ReadBlock8HP(src, read_dst, dstpitch);
	}
	FOREACH_BLOCK_END
}

void GSLocalMemory::ReadTexture4HLP(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	FOREACH_BLOCK_START(r, 8, 8, 8)
	{
		GSBlock::ReadBlock4HLP(src, read_dst, dstpitch);
	}
	FOREACH_BLOCK_END
}

void GSLocalMemory::ReadTexture4HHP(const GSOffset* RESTRICT off, const GSVector4i& r, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA)
{
	FOREACH_BLOCK_START(r, 8, 8, 8)
	{
		GSBlock::ReadBlock4HHP(src, read_dst, dstpitch);
	}
	FOREACH_BLOCK_END
}

//

void GSLocalMemory::ReadTextureBlock8P(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	GSBlock::ReadBlock8(BlockPtr(bp), dst, dstpitch);
}

void GSLocalMemory::ReadTextureBlock4P(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	GSBlock::ReadBlock4P(BlockPtr(bp), dst, dstpitch);
}

void GSLocalMemory::ReadTextureBlock8HP(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	GSBlock::ReadBlock8HP(BlockPtr(bp), dst, dstpitch);
}

void GSLocalMemory::ReadTextureBlock4HLP(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	GSBlock::ReadBlock4HLP(BlockPtr(bp), dst, dstpitch);
}

void GSLocalMemory::ReadTextureBlock4HHP(uint32 bp, uint8* dst, int dstpitch, const GIFRegTEXA& TEXA) const
{
	ALIGN_STACK(32);

	GSBlock::ReadBlock4HHP(BlockPtr(bp), dst, dstpitch);
}

//

#include "Renderers/SW/GSTextureSW.h"

void GSLocalMemory::SaveBMP(const std::string& fn, uint32 bp, uint32 bw, uint32 psm, int w, int h)
{
	int pitch = w * 4;
	int size = pitch * h;
	void* bits = _aligned_malloc(size, 32);

	GIFRegTEX0 TEX0;

	TEX0.TBP0 = bp;
	TEX0.TBW = bw;
	TEX0.PSM = psm;

	readPixel rp = m_psm[psm].rp;

	uint8* p = (uint8*)bits;

	for(int j = 0; j < h; j++, p += pitch)
	{
		for(int i = 0; i < w; i++)
		{
			((uint32*)p)[i] = (this->*rp)(i, j, TEX0.TBP0, TEX0.TBW);
		}
	}

	GSTextureSW t(GSTexture::Offscreen, w, h);

	if(t.Update(GSVector4i(0, 0, w, h), bits, pitch))
	{
		t.Save(fn);
	}

	_aligned_free(bits);
}

// GSOffset

GSOffset::GSOffset(uint32 _bp, uint32 _bw, uint32 _psm)
{
	hash = _bp | (_bw << 14) | (_psm << 20);

	GSLocalMemory::pixelAddress bn = GSLocalMemory::m_psm[_psm].bn;

	for(int i = 0; i < 256; i++)
	{
		block.row[i] = (short)bn(0, i << 3, _bp, _bw);
	}

	block.col = GSLocalMemory::m_psm[_psm].blockOffset;

	GSLocalMemory::pixelAddress pa = GSLocalMemory::m_psm[_psm].pa;

	for(int i = 0; i < 4096; i++)
	{
		pixel.row[i] = (int)pa(0, i & 0x7ff, _bp, _bw);
	}

	for(int i = 0; i < 8; i++)
	{
		pixel.col[i] = GSLocalMemory::m_psm[_psm].rowOffset[i];
	}

	pages_as_bit.fill(nullptr);
}

GSOffset::~GSOffset()
{
	for(auto buffer: pages_as_bit)
		_aligned_free(buffer);
}

uint32* GSOffset::GetPages(const GSVector4i& rect, uint32* pages, GSVector4i* bbox)
{
	GSVector2i bs = (bp & 31) == 0 ? GSLocalMemory::m_psm[psm].pgs : GSLocalMemory::m_psm[psm].bs;

	GSVector4i r = rect.ralign<Align_Outside>(bs);

	if(bbox != NULL) *bbox = r;

	// worst case:
	// bp page-aligned: (w * h) / (64 * 32)
	// bp block-aligned: (w * h) / (8 * 8)

	int size = r.width() * r.height();

	int limit = MAX_PAGES + 1;

	if(pages == NULL)
	{
		limit = std::min<int>((size >> ((bp & 31) != 0 ? 6 : 11)) + 2, MAX_PAGES) + 1;

		pages = new uint32[limit];
	}

	alignas(16) uint32 tmp[16];

	((GSVector4i*)tmp)[0] = GSVector4i::zero();
	((GSVector4i*)tmp)[1] = GSVector4i::zero();
	((GSVector4i*)tmp)[2] = GSVector4i::zero();
	((GSVector4i*)tmp)[3] = GSVector4i::zero();

	r = r.sra32(3);

	bs.x >>= 3;
	bs.y >>= 3;

	uint32* RESTRICT p = pages;

	for(int y = r.top; y < r.bottom; y += bs.y)
	{
		uint32 base = block.row[y];

		for(int x = r.left; x < r.right; x += bs.x)
		{
			uint32 n = ((base + block.col[x]) >> 5) % MAX_PAGES;

			uint32& row = tmp[n >> 5];
			uint32 col = 1 << (n & 31);

			if((row & col) == 0)
			{
				row |= col;

				*p++ = n;
			}
		}
	}

	*p++ = (uint32)EOP;

	ASSERT(p - pages <= limit);

	return pages;
}

uint32* GSOffset::GetPagesAsBits(const GIFRegTEX0& TEX0)
{
	// Performance note:
	// GSOffset is per bp/bw/psm
	// Pages coverage depends also on TW and Th (8bits). Therefore we will save them as a small array.
	// It is faster than a hash cache and it reduces the GetPagesAsBits overhead.

	int hash_key = (TEX0.u64 >> 26) & 0xFF;
	uint32* pages = pages_as_bit[hash_key];

	if (pages)
		return pages;

	// Aligned on 64 bytes to store the full bitmap in a single cache line
	pages = (uint32*)_aligned_malloc(MAX_PAGES/8, 64);
	pages_as_bit[hash_key] = pages;

	GetPagesAsBits(GSVector4i(0, 0, 1 << TEX0.TW, 1 << TEX0.TH), pages);

	return pages;
}


void* GSOffset::GetPagesAsBits(const GSVector4i& rect, void* pages)
{
	ASSERT(pages != nullptr);

	((GSVector4i*)pages)[0] = GSVector4i::zero();
	((GSVector4i*)pages)[1] = GSVector4i::zero();
	((GSVector4i*)pages)[2] = GSVector4i::zero();
	((GSVector4i*)pages)[3] = GSVector4i::zero();

	GSVector2i bs = (bp & 31) == 0 ? GSLocalMemory::m_psm[psm].pgs : GSLocalMemory::m_psm[psm].bs;

	GSVector4i r = rect.ralign<Align_Outside>(bs);

	r = r.sra32(3);

	bs.x >>= 3;
	bs.y >>= 3;

	for(int y = r.top; y < r.bottom; y += bs.y)
	{
		uint32 base = block.row[y];

		for(int x = r.left; x < r.right; x += bs.x)
		{
			uint32 n = ((base + block.col[x]) >> 5) % MAX_PAGES;

			((uint32*)pages)[n >> 5] |= 1 << (n & 31);
		}
	}

	return pages;
}