archive/dolphin
1
0
Fork 0
mirror of https://github.com/dolphin-emu/dolphin.git synced 2024-09-20 11:21:43 +02:00
Code Issues Projects Releases Packages Wiki Activity

revert last 2 commits

Browse source 
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@3533 8ced0084-cf51-0410-be5f-012b33b47a6e
This commit is contained in:
nakeee 2009-06-22 09:31:30 +00:00
parent 5bcf69cbd6
commit 6b2855de5e
37 changed files with 1935 additions and 1939 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

38
Source/Core/VideoCommon/Src/SUFunctions.h → Source/Core/VideoCommon/Src/BPFunctions.h
View file

@ -20,13 +20,13 @@
// The plugins has to define these functions
// ------------------------------------------
#ifndef _SUFUNCTIONS_H_
#define _SUFUNCTIONS_H_
#ifndef _BPFUNCTIONS_H_
#define _BPFUNCTIONS_H_
#include "SUMemory.h"
#include "BPMemory.h"
#include "VideoCommon.h"
namespace SUFunctions
namespace BPFunctions
{
enum
@ -37,24 +37,24 @@ enum
};
void FlushPipeline();
void SetGenerationMode(const BPCommand &bp);
void SetScissor(const BPCommand &bp);
void SetLineWidth(const BPCommand &bp);
void SetDepthMode(const BPCommand &bp);
void SetBlendMode(const BPCommand &bp);
void SetDitherMode(const BPCommand &bp);
void SetLogicOpMode(const BPCommand &bp);
void SetColorMask(const BPCommand &bp);
void SetGenerationMode(const Bypass &bp);
void SetScissor(const Bypass &bp);
void SetLineWidth(const Bypass &bp);
void SetDepthMode(const Bypass &bp);
void SetBlendMode(const Bypass &bp);
void SetDitherMode(const Bypass &bp);
void SetLogicOpMode(const Bypass &bp);
void SetColorMask(const Bypass &bp);
float GetRendererTargetScaleX();
float GetRendererTargetScaleY();
void CopyEFB(const BPCommand &bp, const TRectangle &rc, const u32 &address, const bool &fromZBuffer, const bool &isIntensityFmt, const u32 &copyfmt, const bool &scaleByHalf);
void RenderToXFB(const BPCommand &bp, const TRectangle &multirc, const float &yScale, const float &xfbLines, u8* pXFB, const u32 &dstWidth, const u32 &dstHeight);
void ClearScreen(const BPCommand &bp, const TRectangle &multirc);
void RestoreRenderState(const BPCommand &bp);
void CopyEFB(const Bypass &bp, const TRectangle &rc, const u32 &address, const bool &fromZBuffer, const bool &isIntensityFmt, const u32 &copyfmt, const bool &scaleByHalf);
void RenderToXFB(const Bypass &bp, const TRectangle &multirc, const float &yScale, const float &xfbLines, u8* pXFB, const u32 &dstWidth, const u32 &dstHeight);
void ClearScreen(const Bypass &bp, const TRectangle &multirc);
void RestoreRenderState(const Bypass &bp);
u8 *GetPointer(const u32 &address);
bool GetConfig(const int &type);
void SetSamplerState(const BPCommand &bp);
void SetInterlacingMode(const BPCommand &bp);
void SetSamplerState(const Bypass &bp);
void SetInterlacingMode(const Bypass &bp);
};
#endif // _SUFUNCTIONS_H_
#endif // _BPFUNCTIONS_H_

136
Source/Core/VideoCommon/Src/SUMemory.cpp → Source/Core/VideoCommon/Src/BPMemory.cpp
View file

@ -1,68 +1,68 @@
// Copyright (C) 2003-2008 Dolphin Project.
// 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, version 2.0.
// 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 2.0 for more details.
// A copy of the GPL 2.0 should have been included with the program.
// If not, see http://www.gnu.org/licenses/
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#include "Common.h"
#include "SUMemory.h"
// SU state
// STATE_TO_SAVE
SUMemory sumem;
// The plugin must implement this.
void SUWritten(const BPCommand& bp);
// Call browser: OpcodeDecoding.cpp ExecuteDisplayList > Decode() > LoadSUReg()
void LoadSUReg(u32 value0)
{
//handle the mask register
int opcode = value0 >> 24;
int oldval = ((u32*)&sumem)[opcode];
int newval = (oldval & ~sumem.suMask) | (value0 & sumem.suMask);
int changes = (oldval ^ newval) & 0xFFFFFF;
BPCommand bp = {opcode, changes, newval};
//reset the mask register
if (opcode != 0xFE)
sumem.suMask = 0xFFFFFF;
SUWritten(bp);
}
// Called when loading a saved state.
// Needs more testing though.
void SUReload()
{
for (int i = 0; i < 254; i++)
{
switch (i) {
case 0x41:
case 0x45: //GXSetDrawDone
case 0x52:
case 0x65:
case 0x67: // set gp metric?
case SUMEM_PE_TOKEN_ID:
case SUMEM_PE_TOKEN_INT_ID:
// Cases in which we DON'T want to reload the BP
continue;
default:
BPCommand bp = {i, 0xFFFFFF, ((u32*)&sumem)[i]};
SUWritten(bp);
}
}
}
// Copyright (C) 2003-2008 Dolphin Project.
// 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, version 2.0.
// 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 2.0 for more details.
// A copy of the GPL 2.0 should have been included with the program.
// If not, see http://www.gnu.org/licenses/
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#include "Common.h"
#include "BPMemory.h"
//BP state
// STATE_TO_SAVE
BPMemory bpmem;
// The plugin must implement this.
void BPWritten(const Bypass& bp);
// Call browser: OpcodeDecoding.cpp ExecuteDisplayList > Decode() > LoadBPReg()
void LoadBPReg(u32 value0)
{
//handle the mask register
int opcode = value0 >> 24;
int oldval = ((u32*)&bpmem)[opcode];
int newval = (oldval & ~bpmem.bpMask) | (value0 & bpmem.bpMask);
int changes = (oldval ^ newval) & 0xFFFFFF;
Bypass bp = {opcode, changes, newval};
//reset the mask register
if (opcode != 0xFE)
bpmem.bpMask = 0xFFFFFF;
BPWritten(bp);
}
// Called when loading a saved state.
// Needs more testing though.
void BPReload()
{
for (int i = 0; i < 254; i++)
{
switch (i) {
case 0x41:
case 0x45: //GXSetDrawDone
case 0x52:
case 0x65:
case 0x67: // set gp metric?
case BPMEM_PE_TOKEN_ID:
case BPMEM_PE_TOKEN_INT_ID:
// Cases in which we DON'T want to reload the BP
continue;
default:
Bypass bp = {i, 0xFFFFFF, ((u32*)&bpmem)[i]};
BPWritten(bp);
}
}
}

1827
Source/Core/VideoCommon/Src/SUMemory.h → Source/Core/VideoCommon/Src/BPMemory.h
View file

File diff suppressed because it is too large Load diff

596
Source/Core/VideoCommon/Src/BPStructs.cpp Normal file
View file

@ -0,0 +1,596 @@
// Copyright (C) 2003-2009 Dolphin Project.
// 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, version 2.0.
// 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 2.0 for more details.
// A copy of the GPL 2.0 should have been included with the program.
// If not, see http://www.gnu.org/licenses/
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#include <cmath>
#include "Profiler.h"
#include "Statistics.h"
#include "VideoCommon.h"
#include "PixelShaderManager.h"
#include "BPFunctions.h"
#include "BPStructs.h"
#include "TextureDecoder.h"
#include "OpcodeDecoding.h"
#include "VertexLoader.h"
#include "VertexShaderManager.h"
using namespace BPFunctions;
void BPInit()
{
memset(&bpmem, 0, sizeof(bpmem));
bpmem.bpMask = 0xFFFFFF;
}
// ----------------------------------------------------------------------------------------------------------
// Write to the Bypass Memory (Bypass Raster State Registers)
/* ------------------
Called:
At the end of every: OpcodeDecoding.cpp ExecuteDisplayList > Decode() > LoadBPReg
TODO:
Turn into function table. The (future) DisplayList (DL) jit can then call the functions directly,
getting rid of dynamic dispatch. Unfortunately, few games use DLs properly - most\
just stuff geometry in them and don't put state changes there. */
// ----------------------------------------------------------------------------------------------------------
void BPWritten(const Bypass& bp)
{
// --------------------------------------------------------------------------------------------------------
// First the pipeline is flushed then update the bpmem with the new value.
// Some of the BP cases have to call certain functions while others just update the bpmem.
// some bp cases check the changes variable, because they might not have to be updated all the time
// NOTE: it seems not all bp cases like checking changes, so calling if (bp.changes == 0 ? false : true)
// had to be ditched and the games seem to work fine with out it.
// NOTE2: Yet Another Gamecube Documentation calls them Bypass Registers
// --------------------------------------------------------------------------------------------------------
// Debugging only, this lets you skip a bp update
//static int times = 0;
//static bool enable = false;
//switch (bp.address)
//{
//case BPMEM_CONSTANTALPHA:
// {
// if (times-- == 0 && enable)
// return;
// else
// break;
// }
//default: break;
//}
FlushPipeline();
((u32*)&bpmem)[bp.address] = bp.newvalue;
switch (bp.address)
{
case BPMEM_GENMODE: // Set the Generation Mode
{
PRIM_LOG("genmode: texgen=%d, col=%d, ms_en=%d, tev=%d, culmode=%d, ind=%d, zfeeze=%d",
bpmem.genMode.numtexgens, bpmem.genMode.numcolchans,
bpmem.genMode.ms_en, bpmem.genMode.numtevstages+1, bpmem.genMode.cullmode,
bpmem.genMode.numindstages, bpmem.genMode.zfreeze);
SetGenerationMode(bp);
break;
}
case BPMEM_IND_MTXA: // Index Matrix Changed
case BPMEM_IND_MTXB:
case BPMEM_IND_MTXC:
case BPMEM_IND_MTXA+3:
case BPMEM_IND_MTXB+3:
case BPMEM_IND_MTXC+3:
case BPMEM_IND_MTXA+6:
case BPMEM_IND_MTXB+6:
case BPMEM_IND_MTXC+6:
PixelShaderManager::SetIndMatrixChanged((bp.address - BPMEM_IND_MTXA) / 3);
break;
case BPMEM_RAS1_SS0: // Index Texture Coordinate Scale 0
PixelShaderManager::SetIndTexScaleChanged(0x03);
case BPMEM_RAS1_SS1: // Index Texture Coordinate Scale 1
PixelShaderManager::SetIndTexScaleChanged(0x0c);
break;
case BPMEM_SCISSORTL: // Scissor Rectable Top, Left
case BPMEM_SCISSORBR: // Scissor Rectable Bottom, Right
case BPMEM_SCISSOROFFSET: // Scissor Offset
SetScissor(bp);
break;
case BPMEM_LINEPTWIDTH: // Line Width
SetLineWidth(bp);
break;
case BPMEM_ZMODE: // Depth Control
PRIM_LOG("zmode: test=%d, func=%d, upd=%d", bpmem.zmode.testenable, bpmem.zmode.func,
bpmem.zmode.updateenable);
SetDepthMode(bp);
break;
case BPMEM_BLENDMODE: // Blending Control
{
if (bp.changes & 0xFFFF)
{
PRIM_LOG("blendmode: en=%d, open=%d, colupd=%d, alphaupd=%d, dst=%d, src=%d, sub=%d, mode=%d",
bpmem.blendmode.blendenable, bpmem.blendmode.logicopenable, bpmem.blendmode.colorupdate, bpmem.blendmode.alphaupdate,
bpmem.blendmode.dstfactor, bpmem.blendmode.srcfactor, bpmem.blendmode.subtract, bpmem.blendmode.logicmode);
// Set LogicOp Blending Mode
if (bp.changes & 2)
{
SETSTAT(stats.logicOpMode, bpmem.blendmode.logicopenable != 0 ? bpmem.blendmode.logicmode : stats.logicOpMode);
SetLogicOpMode(bp);
}
// Set Dithering Mode
if (bp.changes & 4)
{
SETSTAT(stats.dither, bpmem.blendmode.dither);
SetDitherMode(bp);
}
// Set Blending Mode
if (bp.changes & 0xFE1)
{
SETSTAT(stats.srcFactor, bpmem.blendmode.srcfactor);
SETSTAT(stats.dstFactor, bpmem.blendmode.dstfactor);
SetBlendMode(bp);
}
// Set Color Mask
if (bp.changes & 0x18)
{
SETSTAT(stats.alphaUpdate, bpmem.blendmode.alphaupdate);
SETSTAT(stats.colorUpdate, bpmem.blendmode.colorupdate);
SetColorMask(bp);
}
}
break;
}
case BPMEM_CONSTANTALPHA: // Set Destination Alpha
{
PRIM_LOG("constalpha: alp=%d, en=%d", bpmem.dstalpha.alpha, bpmem.dstalpha.enable);
SETSTAT(stats.dstAlphaEnable, bpmem.dstalpha.enable);
SETSTAT_UINT(stats.dstAlpha, bpmem.dstalpha.alpha);
PixelShaderManager::SetDestAlpha(bpmem.dstalpha);
break;
}
// This is called when the game is done drawing the new frame (eg: like in DX: Begin(); Draw(); End();)
case BPMEM_SETDRAWDONE:
switch (bp.newvalue & 0xFF)
{
case 0x02:
g_VideoInitialize.pSetPEFinish(); // may generate interrupt
DEBUG_LOG(VIDEO, "GXSetDrawDone SetPEFinish (value: 0x%02X)", (bp.newvalue & 0xFFFF));
break;
default:
WARN_LOG(VIDEO, "GXSetDrawDone ??? (value 0x%02X)", (bp.newvalue & 0xFFFF));
break;
}
break;
case BPMEM_PE_TOKEN_ID: // Pixel Engine Token ID
g_VideoInitialize.pSetPEToken(static_cast<u16>(bp.newvalue & 0xFFFF), FALSE);
DEBUG_LOG(VIDEO, "SetPEToken 0x%04x", (bp.newvalue & 0xFFFF));
break;
case BPMEM_PE_TOKEN_INT_ID: // Pixel Engine Interrupt Token ID
g_VideoInitialize.pSetPEToken(static_cast<u16>(bp.newvalue & 0xFFFF), TRUE);
DEBUG_LOG(VIDEO, "SetPEToken + INT 0x%04x", (bp.newvalue & 0xFFFF));
break;
// ------------------------
// EFB copy command. This copies a rectangle from the EFB to either RAM in a texture format or to XFB as YUYV.
// It can also optionally clear the EFB while copying from it. To emulate this, we of course copy first and clear afterwards.
case BPMEM_TRIGGER_EFB_COPY: // Copy EFB Region or Render to the XFB or Clear the screen.
{
DVSTARTSUBPROFILE("LoadBPReg:swap");
// The bottom right is within the rectangle
// The values in bpmem.copyTexSrcXY and bpmem.copyTexSrcWH are updated in case 0x49 and 0x4a in this function
TRectangle rc = {
(int)(bpmem.copyTexSrcXY.x),
(int)(bpmem.copyTexSrcXY.y),
(int)((bpmem.copyTexSrcXY.x + bpmem.copyTexSrcWH.x + 1)),
(int)((bpmem.copyTexSrcXY.y + bpmem.copyTexSrcWH.y + 1))
};
float MValueX = GetRendererTargetScaleX();
float MValueY = GetRendererTargetScaleY();
// Need another rc here to get it to scale.
// Here the bottom right is the out of the rectangle.
TRectangle multirc = {
(int)(bpmem.copyTexSrcXY.x * MValueX),
(int)(bpmem.copyTexSrcXY.y * MValueY),
(int)((bpmem.copyTexSrcXY.x * MValueX + (bpmem.copyTexSrcWH.x + 1) * MValueX)),
(int)((bpmem.copyTexSrcXY.y * MValueY + (bpmem.copyTexSrcWH.y + 1) * MValueY))
};
UPE_Copy PE_copy;
PE_copy.Hex = bpmem.triggerEFBCopy;
// Check if we are to copy from the EFB or draw to the XFB
if (PE_copy.copy_to_xfb == 0)
{
if (GetConfig(CONFIG_SHOWEFBREGIONS))
stats.efb_regions.push_back(rc);
CopyEFB(bp, rc, bpmem.copyTexDest << 5,
bpmem.zcontrol.pixel_format == PIXELFMT_Z24,
PE_copy.intensity_fmt > 0,
((PE_copy.target_pixel_format / 2) + ((PE_copy.target_pixel_format & 1) * 8)),
PE_copy.half_scale > 0);
}
else
{
// the number of lines copied is determined by the y scale * source efb height
const float yScale = bpmem.dispcopyyscale / 256.0f;
const float xfbLines = ((bpmem.copyTexSrcWH.y + 1.0f) * yScale);
RenderToXFB(bp, multirc, yScale, xfbLines,
Memory_GetPtr(bpmem.copyTexDest << 5),
bpmem.copyMipMapStrideChannels << 4,
(u32)ceil(xfbLines));
}
// Clear the picture after it's done and submitted, to prepare for the next picture
if (PE_copy.clear)
ClearScreen(bp, multirc);
RestoreRenderState(bp);
break;
}
case BPMEM_LOADTLUT0: // Load a Texture Look Up Table
case BPMEM_LOADTLUT1:
{
DVSTARTSUBPROFILE("LoadBPReg:GXLoadTlut");
u32 tlutTMemAddr = (bp.newvalue & 0x3FF) << 9;
u32 tlutXferCount = (bp.newvalue & 0x1FFC00) >> 5;
u8 *ptr = 0;
// TODO - figure out a cleaner way.
if (GetConfig(CONFIG_ISWII))
ptr = GetPointer(bpmem.tlutXferSrc << 5);
else
ptr = GetPointer((bpmem.tlutXferSrc & 0xFFFFF) << 5);
if (ptr)
memcpy_gc(texMem + tlutTMemAddr, ptr, tlutXferCount);
else
PanicAlert("Invalid palette pointer %08x %08x %08x", bpmem.tlutXferSrc, bpmem.tlutXferSrc << 5, (bpmem.tlutXferSrc & 0xFFFFF)<< 5);
// TODO(ector) : kill all textures that use this palette
// Not sure if it's a good idea, though. For now, we hash texture palettes
break;
}
case BPMEM_FOGRANGE: // Fog Settings Control
case BPMEM_FOGPARAM0:
case BPMEM_FOGBMAGNITUDE:
case BPMEM_FOGBEXPONENT:
case BPMEM_FOGPARAM3:
if(!GetConfig(CONFIG_DISABLEFOG))
PixelShaderManager::SetFogParamChanged();
break;
case BPMEM_FOGCOLOR: // Fog Color
PixelShaderManager::SetFogColorChanged();
break;
case BPMEM_ALPHACOMPARE: // Compare Alpha Values
PRIM_LOG("alphacmp: ref0=%d, ref1=%d, comp0=%d, comp1=%d, logic=%d", bpmem.alphaFunc.ref0,
bpmem.alphaFunc.ref1, bpmem.alphaFunc.comp0, bpmem.alphaFunc.comp1, bpmem.alphaFunc.logic);
PixelShaderManager::SetAlpha(bpmem.alphaFunc);
break;
case BPMEM_BIAS: // BIAS
PRIM_LOG("ztex bias=0x%x", bpmem.ztex1.bias);
PixelShaderManager::SetZTextureBias(bpmem.ztex1.bias);
break;
case BPMEM_ZTEX2: // Z Texture type
{
if (bp.changes & 3)
PixelShaderManager::SetZTextureTypeChanged();
#if defined(_DEBUG) || defined(DEBUGFAST)
const char* pzop[] = {"DISABLE", "ADD", "REPLACE", "?"};
const char* pztype[] = {"Z8", "Z16", "Z24", "?"};
PRIM_LOG("ztex op=%s, type=%s", pzop[bpmem.ztex2.op], pztype[bpmem.ztex2.type]);
#endif
break;
}
// ----------------------------------
// Display Copy Filtering Control
// Fields: Destination, Frame2Field, Gamma, Source
// TODO: We might have to implement the gamma one, some games might need this, if they are too dark to see.
// ----------------------------------
case BPMEM_DISPLAYCOPYFILER:
case BPMEM_DISPLAYCOPYFILER+1:
case BPMEM_DISPLAYCOPYFILER+2:
case BPMEM_DISPLAYCOPYFILER+3:
case BPMEM_COPYFILTER0: //GXSetCopyFilter
case BPMEM_COPYFILTER1:
break;
case BPMEM_FIELDMASK: // Interlacing Control
case BPMEM_FIELDMODE:
SetInterlacingMode(bp);
break;
// ---------------------------------------------------
// Debugging/Profiling info, we don't care about them
// ---------------------------------------------------
case BPMEM_CLOCK0: // Some Clock
case BPMEM_CLOCK1: // Some Clock
case BPMEM_SU_COUNTER: // Pixel or Poly Count
case BPMEM_RAS_COUNTER: // Sound Count of something in the Texture Units
case BPMEM_SETGPMETRIC: // Set the Graphic Processor Metric
break;
// ----------------
// EFB Copy config
// ----------------
case BPMEM_EFB_TL: // EFB Source Rect. Top, Left
case BPMEM_EFB_BR: // EFB Source Rect. Bottom, Right (w, h - 1)
case BPMEM_EFB_ADDR: // EFB Target Address
break;
// --------------
// Clear Config
// --------------
case BPMEM_CLEAR_AR: // Alpha and Red Components
case BPMEM_CLEAR_GB: // Green and Blue Components
case BPMEM_CLEAR_Z: // Z Components (24-bit Zbuffer)
break;
// -------------------------
// Culling Occulsion, we don't support this
// let's hope not many games use bboxes..
// TODO(ector): add something that watches bboxes
// -------------------------
case BPMEM_CLEARBBOX1:
case BPMEM_CLEARBBOX2:
break;
case BPMEM_ZCOMPARE: // Set the Z-Compare
case BPMEM_TEXINVALIDATE: // Used, if game has manual control the Texture Cache, which we don't allow
case BPMEM_MIPMAP_STRIDE: // MipMap Stride Channel
case BPMEM_COPYYSCALE: // Display Copy Y Scale
case BPMEM_IREF: /* 24 RID
21 BC3 - Ind. Tex Stage 3 NTexCoord
18 BI3 - Ind. Tex Stage 3 NTexMap
15 BC2 - Ind. Tex Stage 2 NTexCoord
12 BI2 - Ind. Tex Stage 2 NTexMap
9 BC1 - Ind. Tex Stage 1 NTexCoord
6 BI1 - Ind. Tex Stage 1 NTexMap
3 BC0 - Ind. Tex Stage 0 NTexCoord
0 BI0 - Ind. Tex Stage 0 NTexMap */
break;
case BPMEM_TEV_KSEL: // Texture Environment Swap Mode Table 0
case BPMEM_TEV_KSEL+1:// Texture Environment Swap Mode Table 1
case BPMEM_TEV_KSEL+2:// Texture Environment Swap Mode Table 2
case BPMEM_TEV_KSEL+3:// Texture Environment Swap Mode Table 3
case BPMEM_TEV_KSEL+4:// Texture Environment Swap Mode Table 4
case BPMEM_TEV_KSEL+5:// Texture Environment Swap Mode Table 5
case BPMEM_TEV_KSEL+6:// Texture Environment Swap Mode Table 6
case BPMEM_TEV_KSEL+7:// Texture Environment Swap Mode Table 7
break;
case BPMEM_BP_MASK: // This Register can be used to limit to which bits of BP registers is actually written to. the mask is
// only valid for the next BP command, and will reset itself.
case BPMEM_IND_IMASK: // Index Mask ?
break;
case BPMEM_UNKNOWN: // This is always set to 0xF at boot of any game, so this sounds like a useless reg
if (bp.newvalue != 0x0F)
PanicAlert("Unknown is not 0xF! val = 0x%08x", bp.newvalue);
break;
// Cases added due to: http://code.google.com/p/dolphin-emu/issues/detail?id=360#c90
// Are these related to BBox?
case BPMEM_UNKNOWN1:
case BPMEM_UNKNOWN2:
case BPMEM_UNKNOWN3:
case BPMEM_UNKNOWN4:
break;
// ------------------------------------------------
// On Default, we try to look for other things
// before we give up and say its an unknown opcode
// ------------------------------------------------
default:
switch (bp.address & 0xFC) // Texture sampler filter
{
// -------------------------
// Texture Environment Order
// -------------------------
case BPMEM_TREF:
case BPMEM_TREF+1:
case BPMEM_TREF+2:
case BPMEM_TREF+3:
case BPMEM_TREF+4:
case BPMEM_TREF+5:
case BPMEM_TREF+6:
case BPMEM_TREF+7:
break;
// ----------------------
// Set wrap size
// ----------------------
case BPMEM_SU_SSIZE:
case BPMEM_SU_TSIZE:
case BPMEM_SU_SSIZE+2:
case BPMEM_SU_TSIZE+2:
case BPMEM_SU_SSIZE+4:
case BPMEM_SU_TSIZE+4:
case BPMEM_SU_SSIZE+6:
case BPMEM_SU_TSIZE+6:
case BPMEM_SU_SSIZE+8:
case BPMEM_SU_TSIZE+8:
case BPMEM_SU_SSIZE+10:
case BPMEM_SU_TSIZE+10:
case BPMEM_SU_SSIZE+12:
case BPMEM_SU_TSIZE+12:
case BPMEM_SU_SSIZE+14:
case BPMEM_SU_TSIZE+14:
PixelShaderManager::SetTexCoordChanged((bp.address - BPMEM_SU_SSIZE) >> 1);
break;
// ------------------------
// BPMEM_TX_SETMODE0 - (Texture lookup and filtering mode) LOD/BIAS Clamp, MaxAnsio, LODBIAS, DiagLoad, Min Filter, Mag Filter, Wrap T, S
// BPMEM_TX_SETMODE1 - (LOD Stuff) - Max LOD, Min LOD
// ------------------------
case BPMEM_TX_SETMODE0: // (0x90 for linear)
case BPMEM_TX_SETMODE0+1:
case BPMEM_TX_SETMODE0+2:
case BPMEM_TX_SETMODE0+3:
case BPMEM_TX_SETMODE1:
case BPMEM_TX_SETMODE1+1:
case BPMEM_TX_SETMODE1+2:
case BPMEM_TX_SETMODE1+3:
case BPMEM_TX_SETMODE0_4:
case BPMEM_TX_SETMODE0_4+1:
case BPMEM_TX_SETMODE0_4+2:
case BPMEM_TX_SETMODE0_4+3:
case BPMEM_TX_SETMODE1_4:
case BPMEM_TX_SETMODE1_4+1:
case BPMEM_TX_SETMODE1_4+2:
case BPMEM_TX_SETMODE1_4+3:
SetSamplerState(bp);
break;
// --------------------------------------------
// BPMEM_TX_SETIMAGE0 - Texture width, height, format
// BPMEM_TX_SETIMAGE1 - even LOD address in TMEM - Image Type, Cache Height, Cache Width, TMEM Offset
// BPMEM_TX_SETIMAGE2 - odd LOD address in TMEM - Cache Height, Cache Width, TMEM Offset
// BPMEM_TX_SETIMAGE3 - Address of Texture in main memory
// --------------------------------------------
case BPMEM_TX_SETIMAGE0:
case BPMEM_TX_SETIMAGE0+1:
case BPMEM_TX_SETIMAGE0+2:
case BPMEM_TX_SETIMAGE0+3:
case BPMEM_TX_SETIMAGE0_4:
case BPMEM_TX_SETIMAGE0_4+1:
case BPMEM_TX_SETIMAGE0_4+2:
case BPMEM_TX_SETIMAGE0_4+3:
case BPMEM_TX_SETIMAGE1:
case BPMEM_TX_SETIMAGE1+1:
case BPMEM_TX_SETIMAGE1+2:
case BPMEM_TX_SETIMAGE1+3:
case BPMEM_TX_SETIMAGE1_4:
case BPMEM_TX_SETIMAGE1_4+1:
case BPMEM_TX_SETIMAGE1_4+2:
case BPMEM_TX_SETIMAGE1_4+3:
case BPMEM_TX_SETIMAGE2:
case BPMEM_TX_SETIMAGE2+1:
case BPMEM_TX_SETIMAGE2+2:
case BPMEM_TX_SETIMAGE2+3:
case BPMEM_TX_SETIMAGE2_4:
case BPMEM_TX_SETIMAGE2_4+1:
case BPMEM_TX_SETIMAGE2_4+2:
case BPMEM_TX_SETIMAGE2_4+3:
case BPMEM_TX_SETIMAGE3:
case BPMEM_TX_SETIMAGE3+1:
case BPMEM_TX_SETIMAGE3+2:
case BPMEM_TX_SETIMAGE3+3:
case BPMEM_TX_SETIMAGE3_4:
case BPMEM_TX_SETIMAGE3_4+1:
case BPMEM_TX_SETIMAGE3_4+2:
case BPMEM_TX_SETIMAGE3_4+3:
break;
// -------------------------------
// Set a TLUT
// BPMEM_TX_SETTLUT - Format, TMEM Offset (offset of TLUT from start of TMEM high bank > > 5)
// -------------------------------
case BPMEM_TX_SETTLUT:
case BPMEM_TX_SETTLUT+1:
case BPMEM_TX_SETTLUT+2:
case BPMEM_TX_SETTLUT+3:
case BPMEM_TX_SETLUT_4:
case BPMEM_TX_SETLUT_4+1:
case BPMEM_TX_SETLUT_4+2:
case BPMEM_TX_SETLUT_4+3:
break;
// ---------------------------------------------------
// Set the TEV Color
// ---------------------------------------------------
case BPMEM_TEV_REGISTER_L: // Reg 1
case BPMEM_TEV_REGISTER_H:
case BPMEM_TEV_REGISTER_L+2: // Reg 2
case BPMEM_TEV_REGISTER_H+2:
case BPMEM_TEV_REGISTER_L+4: // Reg 3
case BPMEM_TEV_REGISTER_H+4:
case BPMEM_TEV_REGISTER_L+6: // Reg 4
case BPMEM_TEV_REGISTER_H+6:
{
if (bp.address & 1)
{
// don't compare with changes!
int num = (bp.address >> 1 ) & 0x3;
PixelShaderManager::SetColorChanged(bpmem.tevregs[num].high.type, num);
}
break;
}
// ------------------------------------------------
// On Default, we try to look for other things
// before we give up and say its an unknown opcode
// again ...
// ------------------------------------------------
default:
switch (bp.address & 0xF0)
{
// --------------
// Indirect Tev
// --------------
case BPMEM_IND_CMD:
case BPMEM_IND_CMD+1:
case BPMEM_IND_CMD+2:
case BPMEM_IND_CMD+3:
case BPMEM_IND_CMD+4:
case BPMEM_IND_CMD+5:
case BPMEM_IND_CMD+6:
case BPMEM_IND_CMD+7:
case BPMEM_IND_CMD+8:
case BPMEM_IND_CMD+9:
case BPMEM_IND_CMD+10:
case BPMEM_IND_CMD+11:
case BPMEM_IND_CMD+12:
case BPMEM_IND_CMD+13:
case BPMEM_IND_CMD+14:
case BPMEM_IND_CMD+15:
break;
// --------------------------------------------------
// Set Color/Alpha of a Tev
// BPMEM_TEV_COLOR_ENV - Dest, Shift, Clamp, Sub, Bias, Sel A, Sel B, Sel C, Sel D
// BPMEM_TEV_ALPHA_ENV - Dest, Shift, Clamp, Sub, Bias, Sel A, Sel B, Sel C, Sel D, T Swap, R Swap
// --------------------------------------------------
case BPMEM_TEV_COLOR_ENV: // Texture Environment 1
case BPMEM_TEV_ALPHA_ENV:
case BPMEM_TEV_COLOR_ENV+2: // Texture Environment 2
case BPMEM_TEV_ALPHA_ENV+2:
case BPMEM_TEV_COLOR_ENV+4: // Texture Environment 3
case BPMEM_TEV_ALPHA_ENV+4:
case BPMEM_TEV_COLOR_ENV+8: // Texture Environment 4
case BPMEM_TEV_ALPHA_ENV+8:
case BPMEM_TEV_COLOR_ENV+10: // Texture Environment 5
case BPMEM_TEV_ALPHA_ENV+10:
case BPMEM_TEV_COLOR_ENV+12: // Texture Environment 6
case BPMEM_TEV_ALPHA_ENV+12:
case BPMEM_TEV_COLOR_ENV+14: // Texture Environment 7
case BPMEM_TEV_ALPHA_ENV+14:
case BPMEM_TEV_COLOR_ENV+16: // Texture Environment 8
case BPMEM_TEV_ALPHA_ENV+16:
case BPMEM_TEV_COLOR_ENV+18: // Texture Environment 9
case BPMEM_TEV_ALPHA_ENV+18:
case BPMEM_TEV_COLOR_ENV+20: // Texture Environment 10
case BPMEM_TEV_ALPHA_ENV+20:
case BPMEM_TEV_COLOR_ENV+22: // Texture Environment 11
case BPMEM_TEV_ALPHA_ENV+22:
case BPMEM_TEV_COLOR_ENV+24: // Texture Environment 12
case BPMEM_TEV_ALPHA_ENV+24:
case BPMEM_TEV_COLOR_ENV+26: // Texture Environment 13
case BPMEM_TEV_ALPHA_ENV+26:
case BPMEM_TEV_COLOR_ENV+28: // Texture Environment 14
case BPMEM_TEV_ALPHA_ENV+28:
case BPMEM_TEV_COLOR_ENV+30: // Texture Environment 15
case BPMEM_TEV_ALPHA_ENV+30:
case BPMEM_TEV_COLOR_ENV+32: // Texture Environment 16
case BPMEM_TEV_ALPHA_ENV+32:
break;
default:
WARN_LOG(VIDEO, "Unknown Bypass opcode: address = 0x%08x value = 0x%08x", bp.address, bp.newvalue);
break;
}
}
}
}

14
Source/Core/VideoCommon/Src/SUStructs.h → Source/Core/VideoCommon/Src/BPStructs.h
View file

@ -15,13 +15,13 @@
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#ifndef _SUSTRUCTS_H_
#define _SUSTRUCTS_H_
#ifndef _BPSTRUCTS_H_
#define _BPSTRUCTS_H_
#include "SUMemory.h"
#include "BPMemory.h"
void SUInit();
void LoadSUReg(u32 value0);
void SUReload();
void BPInit();
void LoadBPReg(u32 value0);
void BPReload();
#endif // _SUSTRUCTS_H_
#endif // _BPSTRUCTS_H_

4
Source/Core/VideoCommon/Src/OpcodeDecoding.cpp
View file

@ -35,7 +35,7 @@
#include "XFMemory.h"
#include "CPMemory.h"
#include "SUMemory.h"
#include "BPMemory.h"
#include "Fifo.h"
#include "DataReader.h"
@ -267,7 +267,7 @@ static void Decode()
case GX_LOAD_BP_REG: //0x61
{
u32 cmd = DataReadU32();
LoadSUReg(cmd);
LoadBPReg(cmd);
INCSTAT(stats.thisFrame.numBPLoads);
}
break;

194
Source/Core/VideoCommon/Src/PixelShaderGen.cpp
View file

@ -22,63 +22,63 @@
#include "Profiler.h"
#include "PixelShaderGen.h"
#include "XFMemory.h" // for texture projection mode
#include "SUMemory.h"
#include "BPMemory.h"
// Mash together all the inputs that contribute to the code of a generated pixel shader into
// a unique identifier, basically containing all the bits. Yup, it's a lot ....
void GetPixelShaderId(PIXELSHADERUID &uid, u32 s_texturemask, u32 dstAlphaEnable)
{
u32 projtexcoords = 0;
for (u32 i = 0; i < (u32)sumem.genMode.numtevstages + 1; i++) {
if (sumem.tevorders[i/2].getEnable(i&1)) {
int texcoord = sumem.tevorders[i/2].getTexCoord(i&1);
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; i++) {
if (bpmem.tevorders[i/2].getEnable(i&1)) {
int texcoord = bpmem.tevorders[i/2].getTexCoord(i&1);
if (xfregs.texcoords[texcoord].texmtxinfo.projection )
projtexcoords |= 1 << texcoord;
}
}
uid.values[0] = (u32)sumem.genMode.numtevstages |
((u32)sumem.genMode.numindstages << 4) |
((u32)sumem.genMode.numtexgens << 7) |
uid.values[0] = (u32)bpmem.genMode.numtevstages |
((u32)bpmem.genMode.numindstages << 4) |
((u32)bpmem.genMode.numtexgens << 7) |
((u32)dstAlphaEnable << 11) |
((u32)((sumem.alphaFunc.hex >> 16) & 0xff) << 12) |
((u32)((bpmem.alphaFunc.hex >> 16) & 0xff) << 12) |
(projtexcoords << 20) |
((u32)sumem.ztex2.op << 28);
((u32)bpmem.ztex2.op << 28);
uid.values[0] = (uid.values[0] & ~0x0ff00000) | (projtexcoords << 20);
// swap table
for (int i = 0; i < 8; i += 2)
((u8*)&uid.values[1])[i/2] = (sumem.tevksel[i].hex & 0xf) | ((sumem.tevksel[i + 1].hex & 0xf) << 4);
((u8*)&uid.values[1])[i/2] = (bpmem.tevksel[i].hex & 0xf) | ((bpmem.tevksel[i + 1].hex & 0xf) << 4);
uid.values[2] = s_texturemask;
uid.values[3] = (u32)sumem.fog.c_proj_fsel.fsel |
((u32)sumem.fog.c_proj_fsel.proj << 3);
uid.values[3] = (u32)bpmem.fog.c_proj_fsel.fsel |
((u32)bpmem.fog.c_proj_fsel.proj << 3);
int hdr = 4;
u32* pcurvalue = &uid.values[hdr];
for (u32 i = 0; i < (u32)sumem.genMode.numtevstages+1; ++i) {
TevStageCombiner::ColorCombiner &cc = sumem.combiners[i].colorC;
TevStageCombiner::AlphaCombiner &ac = sumem.combiners[i].alphaC;
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages+1; ++i) {
TevStageCombiner::ColorCombiner &cc = bpmem.combiners[i].colorC;
TevStageCombiner::AlphaCombiner &ac = bpmem.combiners[i].alphaC;
u32 val0 = cc.hex&0xffffff;
u32 val1 = ac.hex&0xffffff;
val0 |= sumem.tevksel[i/2].getKC(i&1)<<24;
val1 |= sumem.tevksel[i/2].getKA(i&1)<<24;
val0 |= bpmem.tevksel[i/2].getKC(i&1)<<24;
val1 |= bpmem.tevksel[i/2].getKA(i&1)<<24;
pcurvalue[0] = val0;
pcurvalue[1] = val1;
pcurvalue += 2;
}
for (u32 i = 0; i < ((u32)sumem.genMode.numtevstages+1)/2; ++i) {
for (u32 i = 0; i < ((u32)bpmem.genMode.numtevstages+1)/2; ++i) {
u32 val0, val1;
if (sumem.tevorders[i].hex & 0x40)
val0 = sumem.tevorders[i].hex & 0x3ff;
if (bpmem.tevorders[i].hex & 0x40)
val0 = bpmem.tevorders[i].hex & 0x3ff;
else
val0 = sumem.tevorders[i].hex & 0x380;
if (sumem.tevorders[i].hex & 0x40000)
val1 = (sumem.tevorders[i].hex & 0x3ff000) >> 12;
val0 = bpmem.tevorders[i].hex & 0x380;
if (bpmem.tevorders[i].hex & 0x40000)
val1 = (bpmem.tevorders[i].hex & 0x3ff000) >> 12;
else
val1 = (sumem.tevorders[i].hex & 0x380000) >> 12;
val1 = (bpmem.tevorders[i].hex & 0x380000) >> 12;
switch (i % 3) {
case 0: pcurvalue[0] = val0|(val1<<10); break;
@ -87,27 +87,27 @@ void GetPixelShaderId(PIXELSHADERUID &uid, u32 s_texturemask, u32 dstAlphaEnable
}
}
if ((sumem.genMode.numtevstages + 1) & 1) { // odd
if ((bpmem.genMode.numtevstages + 1) & 1) { // odd
u32 val0;
if (sumem.tevorders[sumem.genMode.numtevstages/2].hex & 0x40)
val0 = sumem.tevorders[sumem.genMode.numtevstages/2].hex&0x3ff;
if (bpmem.tevorders[bpmem.genMode.numtevstages/2].hex & 0x40)
val0 = bpmem.tevorders[bpmem.genMode.numtevstages/2].hex&0x3ff;
else
val0 = sumem.tevorders[sumem.genMode.numtevstages/2].hex & 0x380;
val0 = bpmem.tevorders[bpmem.genMode.numtevstages/2].hex & 0x380;
switch (sumem.genMode.numtevstages % 3) {
switch (bpmem.genMode.numtevstages % 3) {
case 0: pcurvalue[0] = val0; break;
case 1: pcurvalue[0] |= val0 << 20; break;
case 2: pcurvalue[1] |= val0 << 10; pcurvalue++; break;
}
}
if ((sumem.genMode.numtevstages % 3) != 2)
if ((bpmem.genMode.numtevstages % 3) != 2)
++pcurvalue;
uid.tevstages = (u32)(pcurvalue-&uid.values[0]-hdr);
for (u32 i = 0; i < sumem.genMode.numindstages; ++i) {
u32 val = sumem.tevind[i].hex & 0x1fffff; // 21 bits
for (u32 i = 0; i < bpmem.genMode.numindstages; ++i) {
u32 val = bpmem.tevind[i].hex & 0x1fffff; // 21 bits
switch (i%3) {
case 0: pcurvalue[0] = val; break;
case 1: pcurvalue[0] |= val << 21; pcurvalue[1] = val >> 11; ++pcurvalue; break;
@ -121,9 +121,9 @@ void GetPixelShaderId(PIXELSHADERUID &uid, u32 s_texturemask, u32 dstAlphaEnable
// old tev->pixelshader notes
//
// color for this stage (alpha, color) is given by sumem.tevorders[0].colorchan0
// konstant for this stage (alpha, color) is given by sumem.tevksel
// inputs are given by sumem.combiners[0].colorC.a/b/c/d << could be current chan color
// color for this stage (alpha, color) is given by bpmem.tevorders[0].colorchan0
// konstant for this stage (alpha, color) is given by bpmem.tevksel
// inputs are given by bpmem.combiners[0].colorC.a/b/c/d << could be current chan color
// according to GXTevColorArg table above
// output is given by .outreg
// tevtemp is set according to swapmodetables and
@ -356,13 +356,13 @@ static char text[16384];
static void BuildSwapModeTable()
{
//sumem.tevregs[0].
//bpmem.tevregs[0].
for (int i = 0; i < 4; i++)
{
swapModeTable[i][0] = swapColors[sumem.tevksel[i*2].swap1];
swapModeTable[i][1] = swapColors[sumem.tevksel[i*2].swap2];
swapModeTable[i][2] = swapColors[sumem.tevksel[i*2+1].swap1];
swapModeTable[i][3] = swapColors[sumem.tevksel[i*2+1].swap2];
swapModeTable[i][0] = swapColors[bpmem.tevksel[i*2].swap1];
swapModeTable[i][1] = swapColors[bpmem.tevksel[i*2].swap2];
swapModeTable[i][2] = swapColors[bpmem.tevksel[i*2+1].swap1];
swapModeTable[i][3] = swapColors[bpmem.tevksel[i*2+1].swap2];
swapModeTable[i][4] = 0;
}
}
@ -373,19 +373,19 @@ const char *GeneratePixelShader(u32 texture_mask, bool dstAlphaEnable, bool HLSL
DVSTARTPROFILE();
BuildSwapModeTable();
int numStages = sumem.genMode.numtevstages + 1;
int numTexgen = sumem.genMode.numtexgens;
int numStages = bpmem.genMode.numtevstages + 1;
int numTexgen = bpmem.genMode.numtexgens;
char *p = text;
WRITE(p, "//Pixel Shader for TEV stages\n");
WRITE(p, "//%i TEV stages, %i texgens, %i IND stages\n",
numStages, numTexgen, sumem.genMode.numindstages);
numStages, numTexgen, bpmem.genMode.numindstages);
int nIndirectStagesUsed = 0;
if (sumem.genMode.numindstages > 0) {
if (bpmem.genMode.numindstages > 0) {
for (int i = 0; i < numStages; ++i) {
if (sumem.tevind[i].IsActive() && sumem.tevind[i].bt < sumem.genMode.numindstages) {
nIndirectStagesUsed |= 1<<sumem.tevind[i].bt;
if (bpmem.tevind[i].IsActive() && bpmem.tevind[i].bt < bpmem.genMode.numindstages) {
nIndirectStagesUsed |= 1<<bpmem.tevind[i].bt;
}
}
}
@ -462,9 +462,9 @@ const char *GeneratePixelShader(u32 texture_mask, bool dstAlphaEnable, bool HLSL
}
// indirect texture map lookup
for(u32 i = 0; i < sumem.genMode.numindstages; ++i) {
for(u32 i = 0; i < bpmem.genMode.numindstages; ++i) {
if (nIndirectStagesUsed & (1<<i)) {
int texcoord = sumem.tevindref.getTexCoord(i);
int texcoord = bpmem.tevindref.getTexCoord(i);
if (texcoord < numTexgen) {
WRITE(p, "tempcoord=uv%d.xy * "I_INDTEXSCALE"[%d].%s;\n", texcoord, i/2, (i&1)?"zw":"xy");
@ -475,7 +475,7 @@ const char *GeneratePixelShader(u32 texture_mask, bool dstAlphaEnable, bool HLSL
char buffer[32];
sprintf(buffer, "float3 indtex%d", i);
SampleTexture(p, buffer, "tempcoord", "abg", sumem.tevindref.getTexMap(i), texture_mask);
SampleTexture(p, buffer, "tempcoord", "abg", bpmem.tevindref.getTexMap(i), texture_mask);
}
}
@ -495,17 +495,17 @@ const char *GeneratePixelShader(u32 texture_mask, bool dstAlphaEnable, bool HLSL
WRITE(p, "float zCoord = "I_ZBIAS"[1].x + (clipPos.z / clipPos.w) * "I_ZBIAS"[1].y;\n");
// use the texture input of the last texture stage (textemp), hopefully this has been read and is in correct format...
if (sumem.ztex2.op == ZTEXTURE_ADD) {
if (bpmem.ztex2.op == ZTEXTURE_ADD) {
WRITE(p, "depth = frac(dot("I_ZBIAS"[0].xyzw, textemp.xyzw) + "I_ZBIAS"[1].w + zCoord);\n");
}
else if (sumem.ztex2.op == ZTEXTURE_REPLACE) {
else if (bpmem.ztex2.op == ZTEXTURE_REPLACE) {
WRITE(p, "depth = frac(dot("I_ZBIAS"[0].xyzw, textemp.xyzw) + "I_ZBIAS"[1].w);\n");
}
else {
WRITE(p, "depth = zCoord;\n");
}
//if (sumem.genMode.numindstages ) WRITE(p, "prev.rg = indtex0.xy;\nprev.b = 0;\n");
//if (bpmem.genMode.numindstages ) WRITE(p, "prev.rg = indtex0.xy;\nprev.b = 0;\n");
if (!WriteAlphaTest(p, HLSL)) {
// alpha test will always fail, so restart the shader and just make it an empty function
@ -530,13 +530,13 @@ const char *GeneratePixelShader(u32 texture_mask, bool dstAlphaEnable, bool HLSL
static void WriteStage(char *&p, int n, u32 texture_mask)
{
char *rasswap = swapModeTable[sumem.combiners[n].alphaC.rswap];
char *texswap = swapModeTable[sumem.combiners[n].alphaC.tswap];
char *rasswap = swapModeTable[bpmem.combiners[n].alphaC.rswap];
char *texswap = swapModeTable[bpmem.combiners[n].alphaC.tswap];
int texcoord = sumem.tevorders[n/2].getTexCoord(n&1);
bool bHasTexCoord = (u32)texcoord < sumem.genMode.numtexgens;
bool bHasIndStage = sumem.tevind[n].IsActive() && sumem.tevind[n].bt < sumem.genMode.numindstages;
int texcoord = bpmem.tevorders[n/2].getTexCoord(n&1);
bool bHasTexCoord = (u32)texcoord < bpmem.genMode.numtexgens;
bool bHasIndStage = bpmem.tevind[n].IsActive() && bpmem.tevind[n].bt < bpmem.genMode.numindstages;
// HACK to handle cases where the tex gen is not enabled
if (!bHasTexCoord) {
@ -545,46 +545,46 @@ static void WriteStage(char *&p, int n, u32 texture_mask)
if (bHasIndStage) {
// perform the indirect op on the incoming regular coordinates using indtex%d as the offset coords
if (sumem.tevind[n].bs != ITBA_OFF) {
if (bpmem.tevind[n].bs != ITBA_OFF) {
// write the bump alpha
if (sumem.tevind[n].fmt == ITF_8) {
WRITE(p, "alphabump = indtex%d.%s %s;\n", sumem.tevind[n].bt,
tevIndAlphaSel[sumem.tevind[n].bs], tevIndAlphaScale[sumem.tevind[n].fmt]);
if (bpmem.tevind[n].fmt == ITF_8) {
WRITE(p, "alphabump = indtex%d.%s %s;\n", bpmem.tevind[n].bt,
tevIndAlphaSel[bpmem.tevind[n].bs], tevIndAlphaScale[bpmem.tevind[n].fmt]);
}
else {
// donkopunchstania: really bad way to do this
// cannot always use fract because fract(1.0) is 0.0 when it needs to be 1.0
// omitting fract seems to work as well
WRITE(p, "if (indtex%d.%s >= 1.0f )\n", sumem.tevind[n].bt,
tevIndAlphaSel[sumem.tevind[n].bs]);
WRITE(p, "if (indtex%d.%s >= 1.0f )\n", bpmem.tevind[n].bt,
tevIndAlphaSel[bpmem.tevind[n].bs]);
WRITE(p, " alphabump = 1.0f;\n");
WRITE(p, "else\n");
WRITE(p, " alphabump = fract ( indtex%d.%s %s );\n", sumem.tevind[n].bt,
tevIndAlphaSel[sumem.tevind[n].bs], tevIndAlphaScale[sumem.tevind[n].fmt]);
WRITE(p, " alphabump = fract ( indtex%d.%s %s );\n", bpmem.tevind[n].bt,
tevIndAlphaSel[bpmem.tevind[n].bs], tevIndAlphaScale[bpmem.tevind[n].fmt]);
}
}
// format
WRITE(p, "float3 indtevcrd%d = indtex%d * %s;\n", n, sumem.tevind[n].bt, tevIndFmtScale[sumem.tevind[n].fmt]);
WRITE(p, "float3 indtevcrd%d = indtex%d * %s;\n", n, bpmem.tevind[n].bt, tevIndFmtScale[bpmem.tevind[n].fmt]);
// bias
if (sumem.tevind[n].bias != ITB_NONE )
WRITE(p, "indtevcrd%d.%s += %s;\n", n, tevIndBiasField[sumem.tevind[n].bias], tevIndBiasAdd[sumem.tevind[n].fmt]);
if (bpmem.tevind[n].bias != ITB_NONE )
WRITE(p, "indtevcrd%d.%s += %s;\n", n, tevIndBiasField[bpmem.tevind[n].bias], tevIndBiasAdd[bpmem.tevind[n].fmt]);
// multiply by offset matrix and scale
if (sumem.tevind[n].mid != 0) {
if (sumem.tevind[n].mid <= 3) {
int mtxidx = 2*(sumem.tevind[n].mid-1);
if (bpmem.tevind[n].mid != 0) {
if (bpmem.tevind[n].mid <= 3) {
int mtxidx = 2*(bpmem.tevind[n].mid-1);
WRITE(p, "float2 indtevtrans%d = float2(dot("I_INDTEXMTX"[%d].xyz, indtevcrd%d), dot("I_INDTEXMTX"[%d].xyz, indtevcrd%d));\n",
n, mtxidx, n, mtxidx+1, n);
}
else if (sumem.tevind[n].mid <= 7 && bHasTexCoord) { // s matrix
int mtxidx = 2*(sumem.tevind[n].mid-5);
else if (bpmem.tevind[n].mid <= 7 && bHasTexCoord) { // s matrix
int mtxidx = 2*(bpmem.tevind[n].mid-5);
WRITE(p, "float2 indtevtrans%d = "I_INDTEXMTX"[%d].ww * uv%d.xy * indtevcrd%d.xx;\n", n, mtxidx, texcoord, n);
}
else if (sumem.tevind[n].mid <= 11 && bHasTexCoord) { // t matrix
int mtxidx = 2*(sumem.tevind[n].mid-9);
else if (bpmem.tevind[n].mid <= 11 && bHasTexCoord) { // t matrix
int mtxidx = 2*(bpmem.tevind[n].mid-9);
WRITE(p, "float2 indtevtrans%d = "I_INDTEXMTX"[%d].ww * uv%d.xy * indtevcrd%d.yy;\n", n, mtxidx, texcoord, n);
}
else {
@ -598,28 +598,28 @@ static void WriteStage(char *&p, int n, u32 texture_mask)
// wrapping
// wrap S
if (sumem.tevind[n].sw == ITW_OFF) {
if (bpmem.tevind[n].sw == ITW_OFF) {
WRITE(p, "wrappedcoord.x = uv%d.x;\n", texcoord);
}
else if (sumem.tevind[n].sw == ITW_0) {
else if (bpmem.tevind[n].sw == ITW_0) {
WRITE(p, "wrappedcoord.x = 0.0f;\n");
}
else {
WRITE(p, "wrappedcoord.x = fmod( uv%d.x, %s );\n", texcoord, tevIndWrapStart[sumem.tevind[n].sw]);
WRITE(p, "wrappedcoord.x = fmod( uv%d.x, %s );\n", texcoord, tevIndWrapStart[bpmem.tevind[n].sw]);
}
// wrap T
if (sumem.tevind[n].tw == ITW_OFF) {
if (bpmem.tevind[n].tw == ITW_OFF) {
WRITE(p, "wrappedcoord.y = uv%d.y;\n", texcoord);
}
else if (sumem.tevind[n].tw == ITW_0) {
else if (bpmem.tevind[n].tw == ITW_0) {
WRITE(p, "wrappedcoord.y = 0.0f;\n");
}
else {
WRITE(p, "wrappedcoord.y = fmod( uv%d.y, %s );\n", texcoord, tevIndWrapStart[sumem.tevind[n].tw]);
WRITE(p, "wrappedcoord.y = fmod( uv%d.y, %s );\n", texcoord, tevIndWrapStart[bpmem.tevind[n].tw]);
}
if (sumem.tevind[n].fb_addprev) {
if (bpmem.tevind[n].fb_addprev) {
// add previous tevcoord
WRITE(p, "tevcoord.xy += wrappedcoord + indtevtrans%d;\n", n);
}
@ -628,10 +628,10 @@ static void WriteStage(char *&p, int n, u32 texture_mask)
}
}
WRITE(p, "rastemp=%s.%s;\n",tevRasTable[sumem.tevorders[n/2].getColorChan(n&1)],rasswap);
WRITE(p, "rastemp=%s.%s;\n",tevRasTable[bpmem.tevorders[n/2].getColorChan(n&1)],rasswap);
if (sumem.tevorders[n/2].getEnable(n&1)) {
int texmap = sumem.tevorders[n/2].getTexMap(n&1);
if (bpmem.tevorders[n/2].getEnable(n&1)) {
int texmap = bpmem.tevorders[n/2].getTexMap(n&1);
if(!bHasIndStage) {
// calc tevcord
if(bHasTexCoord) {
@ -646,11 +646,11 @@ static void WriteStage(char *&p, int n, u32 texture_mask)
else
WRITE(p, "textemp=float4(1,1,1,1);\n");
int kc = sumem.tevksel[n/2].getKC(n&1);
int ka = sumem.tevksel[n/2].getKA(n&1);
int kc = bpmem.tevksel[n/2].getKC(n&1);
int ka = bpmem.tevksel[n/2].getKA(n&1);
TevStageCombiner::ColorCombiner &cc = sumem.combiners[n].colorC;
TevStageCombiner::AlphaCombiner &ac = sumem.combiners[n].alphaC;
TevStageCombiner::ColorCombiner &cc = bpmem.combiners[n].colorC;
TevStageCombiner::AlphaCombiner &ac = bpmem.combiners[n].alphaC;
bool bCKonst = cc.a == TEVCOLORARG_KONST || cc.b == TEVCOLORARG_KONST || cc.c == TEVCOLORARG_KONST || cc.d == TEVCOLORARG_KONST;
bool bAKonst = ac.a == TEVALPHAARG_KONST || ac.b == TEVALPHAARG_KONST || ac.c == TEVALPHAARG_KONST || ac.d == TEVALPHAARG_KONST;
@ -797,8 +797,8 @@ static void WriteAlphaCompare(char *&p, int num, int comp)
static bool WriteAlphaTest(char *&p, bool HLSL)
{
u32 op = sumem.alphaFunc.logic;
u32 comp[2] = {sumem.alphaFunc.comp0,sumem.alphaFunc.comp1};
u32 op = bpmem.alphaFunc.logic;
u32 comp[2] = {bpmem.alphaFunc.comp0,bpmem.alphaFunc.comp1};
//first kill all the simple cases
switch(op) {
@ -839,26 +839,26 @@ static bool WriteAlphaTest(char *&p, bool HLSL)
else {
WRITE(p, "discard( ");
}
WriteAlphaCompare(p, 0, sumem.alphaFunc.comp0);
WriteAlphaCompare(p, 0, bpmem.alphaFunc.comp0);
// negated because testing the inverse condition
switch (sumem.alphaFunc.logic) {
switch (bpmem.alphaFunc.logic) {
case 0: WRITE(p, " || "); break; // and
case 1: WRITE(p, " && "); break; // or
case 2: WRITE(p, " == "); break; // xor
case 3: WRITE(p, " != "); break; // xnor
}
WriteAlphaCompare(p, 1, sumem.alphaFunc.comp1);
WriteAlphaCompare(p, 1, bpmem.alphaFunc.comp1);
WRITE(p, ");\n");
return true;
}
static void WriteFog(char *&p)
{
bool enabled = sumem.fog.c_proj_fsel.fsel == 0 ? false : true;
bool enabled = bpmem.fog.c_proj_fsel.fsel == 0 ? false : true;
if (enabled) {
if (sumem.fog.c_proj_fsel.proj == 0) {
if (bpmem.fog.c_proj_fsel.proj == 0) {
// perspective
// ze = A/(B - Zs)
WRITE (p, " float ze = "I_FOG"[1].x / ("I_FOG"[1].y - depth);\n");
@ -871,7 +871,7 @@ static void WriteFog(char *&p)
WRITE (p, " float fog = clamp(ze - "I_FOG"[1].z, 0.0f, 1.0f);\n");
}
switch (sumem.fog.c_proj_fsel.fsel) {
switch (bpmem.fog.c_proj_fsel.fsel) {
case 2: // linear
// empty
break;

54
Source/Core/VideoCommon/Src/PixelShaderManager.cpp
View file

@ -99,7 +99,7 @@ void PixelShaderManager::SetConstants()
{
static float ffrac = 255.0f/256.0f;
float ftemp[4];
switch (sumem.ztex2.type)
switch (bpmem.ztex2.type)
{
case 0:
// 8 bits
@ -122,7 +122,7 @@ void PixelShaderManager::SetConstants()
if (s_bZBiasChanged || s_bDepthRangeChanged)
{
//ERROR_LOG("pixel=%x,%x, bias=%x\n", sumem.zcontrol.pixel_format, sumem.ztex2.type, lastZBias);
//ERROR_LOG("pixel=%x,%x, bias=%x\n", bpmem.zcontrol.pixel_format, bpmem.ztex2.type, lastZBias);
SetPSConstant4f(C_ZBIAS+1, lastDepthRange[0] / 16777216.0f, lastDepthRange[1] / 16777216.0f, 0, (float)( (((int)lastZBias<<8)>>8))/16777216.0f);
s_bZBiasChanged = s_bDepthRangeChanged = false;
}
@ -137,8 +137,8 @@ void PixelShaderManager::SetConstants()
{
for (u32 i = 0; i < 2; ++i)
{
f[2 * i] = sumem.texscale[0].getScaleS(i & 1);
f[2 * i + 1] = sumem.texscale[0].getScaleT(i & 1);
f[2 * i] = bpmem.texscale[0].getScaleS(i & 1);
f[2 * i + 1] = bpmem.texscale[0].getScaleT(i & 1);
PRIM_LOG("tex indscale%d: %f %f\n", i, f[2 * i], f[2 * i + 1]);
}
SetPSConstant4fv(C_INDTEXSCALE, f);
@ -146,8 +146,8 @@ void PixelShaderManager::SetConstants()
if (s_nIndTexScaleChanged & 0x0c) {
for (u32 i = 2; i < 4; ++i) {
f[2 * i] = sumem.texscale[1].getScaleS(i & 1);
f[2 * i + 1] = sumem.texscale[1].getScaleT(i & 1);
f[2 * i] = bpmem.texscale[1].getScaleS(i & 1);
f[2 * i + 1] = bpmem.texscale[1].getScaleT(i & 1);
PRIM_LOG("tex indscale%d: %f %f\n", i, f[2 * i], f[2 * i + 1]);
}
SetPSConstant4fv(C_INDTEXSCALE+1, &f[4]);
@ -162,28 +162,28 @@ void PixelShaderManager::SetConstants()
{
if (s_nIndTexMtxChanged & (1 << i))
{
int scale = ((u32)sumem.indmtx[i].col0.s0 << 0) |
((u32)sumem.indmtx[i].col1.s1 << 2) |
((u32)sumem.indmtx[i].col2.s2 << 4);
int scale = ((u32)bpmem.indmtx[i].col0.s0 << 0) |
((u32)bpmem.indmtx[i].col1.s1 << 2) |
((u32)bpmem.indmtx[i].col2.s2 << 4);
float fscale = powf(2.0f, (float)(scale - 17)) / 1024.0f;
// xyz - static matrix
// TODO w - dynamic matrix scale / 256...... somehow / 4 works better
// rev 2972 - now using / 256.... verify that this works
SetPSConstant4f(C_INDTEXMTX + 2 * i,
sumem.indmtx[i].col0.ma * fscale,
sumem.indmtx[i].col1.mc * fscale,
sumem.indmtx[i].col2.me * fscale,
bpmem.indmtx[i].col0.ma * fscale,
bpmem.indmtx[i].col1.mc * fscale,
bpmem.indmtx[i].col2.me * fscale,
fscale * 4.0f);
SetPSConstant4f(C_INDTEXMTX + 2 * i + 1,
sumem.indmtx[i].col0.mb * fscale,
sumem.indmtx[i].col1.md * fscale,
sumem.indmtx[i].col2.mf * fscale,
bpmem.indmtx[i].col0.mb * fscale,
bpmem.indmtx[i].col1.md * fscale,
bpmem.indmtx[i].col2.mf * fscale,
fscale * 4.0f);
PRIM_LOG("indmtx%d: scale=%f, mat=(%f %f %f; %f %f %f)\n", i,
1024.0f*fscale, sumem.indmtx[i].col0.ma * fscale, sumem.indmtx[i].col1.mc * fscale, sumem.indmtx[i].col2.me * fscale,
sumem.indmtx[i].col0.mb * fscale, sumem.indmtx[i].col1.md * fscale, sumem.indmtx[i].col2.mf * fscale, fscale);
1024.0f*fscale, bpmem.indmtx[i].col0.ma * fscale, bpmem.indmtx[i].col1.mc * fscale, bpmem.indmtx[i].col2.me * fscale,
bpmem.indmtx[i].col0.mb * fscale, bpmem.indmtx[i].col1.md * fscale, bpmem.indmtx[i].col2.mf * fscale, fscale);
}
}
s_nIndTexMtxChanged = 0;
@ -191,15 +191,15 @@ void PixelShaderManager::SetConstants()
if (s_bFogColorChanged)
{
SetPSConstant4f(C_FOG, sumem.fog.color.r / 255.0f, sumem.fog.color.g / 255.0f, sumem.fog.color.b / 255.0f, 0);
SetPSConstant4f(C_FOG, bpmem.fog.color.r / 255.0f, bpmem.fog.color.g / 255.0f, bpmem.fog.color.b / 255.0f, 0);
s_bFogColorChanged = false;
}
if (s_bFogParamChanged)
{
float a = sumem.fog.a.GetA() * ((float)(1 << sumem.fog.b_shift));
float b = ((float)sumem.fog.b_magnitude / 8388638) * ((float)(1 << (sumem.fog.b_shift - 1)));
SetPSConstant4f(C_FOG + 1, a, b, sumem.fog.c_proj_fsel.GetC(), 0);
float a = bpmem.fog.a.GetA() * ((float)(1 << bpmem.fog.b_shift));
float b = ((float)bpmem.fog.b_magnitude / 8388638) * ((float)(1 << (bpmem.fog.b_shift - 1)));
SetPSConstant4f(C_FOG + 1, a, b, bpmem.fog.c_proj_fsel.GetC(), 0);
s_bFogParamChanged = false;
}
for (int i = 0; i < 8; i++)
@ -214,7 +214,7 @@ void PixelShaderManager::SetPSTextureDims(int texid)
float fdims[4];
if (s_texturemask & (1 << texid))
{
TCoordInfo& tc = sumem.texcoords[texid];
TCoordInfo& tc = bpmem.texcoords[texid];
fdims[0] = (float)(lastTexDims[texid] & 0xffff);
fdims[1] = (float)((lastTexDims[texid] >> 16) & 0xfff);
fdims[2] = (float)(tc.s.scale_minus_1 + 1)*lastCustomTexScale[texid][0];
@ -222,7 +222,7 @@ void PixelShaderManager::SetPSTextureDims(int texid)
}
else
{
TCoordInfo& tc = sumem.texcoords[texid];
TCoordInfo& tc = bpmem.texcoords[texid];
fdims[0] = 1.0f / (float)(lastTexDims[texid] & 0xffff);
fdims[1] = 1.0f / (float)((lastTexDims[texid] >> 16) & 0xfff);
fdims[2] = (float)(tc.s.scale_minus_1 + 1) * lastCustomTexScale[texid][0];
@ -235,10 +235,10 @@ void PixelShaderManager::SetPSTextureDims(int texid)
void PixelShaderManager::SetColorChanged(int type, int num)
{
int r = sumem.tevregs[num].low.a;
int a = sumem.tevregs[num].low.b;
int b = sumem.tevregs[num].high.a;
int g = sumem.tevregs[num].high.b;
int r = bpmem.tevregs[num].low.a;
int a = bpmem.tevregs[num].low.b;
int b = bpmem.tevregs[num].high.a;
int g = bpmem.tevregs[num].high.b;
float *pf = &lastRGBAfull[type][num][0];
pf[0] = (float)r / 255.0f;
pf[1] = (float)g / 255.0f;

2
Source/Core/VideoCommon/Src/PixelShaderManager.h
View file

@ -18,7 +18,7 @@
#ifndef _PIXELSHADERMANAGER_H
#define _PIXELSHADERMANAGER_H
#include "SUMemory.h"
#include "BPMemory.h"
#include "PixelShaderGen.h"
void SetPSConstant4f(int const_number, float f1, float f2, float f3, float f4);

4
Source/Core/VideoCommon/Src/SConscript
View file

@ -3,11 +3,11 @@
Import('env')
files = [
'SUMemory.cpp',
'BPMemory.cpp',
'CPMemory.cpp',
'XFMemory.cpp',
'XFStructs.cpp',
'SUStructs.cpp',
'BPStructs.cpp',
'memcpy_amd.cpp',
'LookUpTables.cpp',
'OpcodeDecoding.cpp',

595
Source/Core/VideoCommon/Src/SUStructs.cpp
View file

@ -1,595 +0,0 @@
// Copyright (C) 2003-2009 Dolphin Project.
// 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, version 2.0.
// 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 2.0 for more details.
// A copy of the GPL 2.0 should have been included with the program.
// If not, see http://www.gnu.org/licenses/
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#include <cmath>
#include "Profiler.h"
#include "Statistics.h"
#include "VideoCommon.h"
#include "PixelShaderManager.h"
#include "SUFunctions.h"
#include "SUStructs.h"
#include "TextureDecoder.h"
#include "OpcodeDecoding.h"
#include "VertexLoader.h"
#include "VertexShaderManager.h"
using namespace SUFunctions;
void SUInit()
{
memset(&sumem, 0, sizeof(sumem));
sumem.suMask = 0xFFFFFF;
}
// ----------------------------------------------------------------------------------------------------------
// Write the bypass command to the Setup/Rasterization Memory
/* ------------------
Called:
At the end of every: OpcodeDecoding.cpp ExecuteDisplayList > Decode() > LoadSUReg
TODO:
Turn into function table. The (future) DisplayList (DL) jit can then call the functions directly,
getting rid of dynamic dispatch. Unfortunately, few games use DLs properly - most\
just stuff geometry in them and don't put state changes there. */
// ----------------------------------------------------------------------------------------------------------
void SUWritten(const BPCommand& bp)
{
// --------------------------------------------------------------------------------------------------------
// First the pipeline is flushed then update the sumem with the new value.
// Some of the BP cases have to call certain functions while others just update the sumem.
// some bp cases check the changes variable, because they might not have to be updated all the time
// NOTE: it seems not all su cases like checking changes, so calling if (bp.changes == 0 ? false : true)
// had to be ditched and the games seem to work fine with out it.
// --------------------------------------------------------------------------------------------------------
// Debugging only, this lets you skip a bp update
//static int times = 0;
//static bool enable = false;
//switch (bp.address)
//{
//case SUMEM_CONSTANTALPHA:
// {
// if (times-- == 0 && enable)
// return;
// else
// break;
// }
//default: break;
//}
FlushPipeline();
((u32*)&sumem)[bp.address] = bp.newvalue;
switch (bp.address)
{
case SUMEM_GENMODE: // Set the Generation Mode
{
PRIM_LOG("genmode: texgen=%d, col=%d, ms_en=%d, tev=%d, culmode=%d, ind=%d, zfeeze=%d",
sumem.genMode.numtexgens, sumem.genMode.numcolchans,
sumem.genMode.ms_en, sumem.genMode.numtevstages+1, sumem.genMode.cullmode,
sumem.genMode.numindstages, sumem.genMode.zfreeze);
SetGenerationMode(bp);
break;
}
case SUMEM_IND_MTXA: // Index Matrix Changed
case SUMEM_IND_MTXB:
case SUMEM_IND_MTXC:
case SUMEM_IND_MTXA+3:
case SUMEM_IND_MTXB+3:
case SUMEM_IND_MTXC+3:
case SUMEM_IND_MTXA+6:
case SUMEM_IND_MTXB+6:
case SUMEM_IND_MTXC+6:
PixelShaderManager::SetIndMatrixChanged((bp.address - SUMEM_IND_MTXA) / 3);
break;
case SUMEM_RAS1_SS0: // Index Texture Coordinate Scale 0
PixelShaderManager::SetIndTexScaleChanged(0x03);
case SUMEM_RAS1_SS1: // Index Texture Coordinate Scale 1
PixelShaderManager::SetIndTexScaleChanged(0x0c);
break;
case SUMEM_SCISSORTL: // Scissor Rectable Top, Left
case SUMEM_SCISSORBR: // Scissor Rectable Bottom, Right
case SUMEM_SCISSOROFFSET: // Scissor Offset
SetScissor(bp);
break;
case SUMEM_LINEPTWIDTH: // Line Width
SetLineWidth(bp);
break;
case SUMEM_ZMODE: // Depth Control
PRIM_LOG("zmode: test=%d, func=%d, upd=%d", sumem.zmode.testenable, sumem.zmode.func,
sumem.zmode.updateenable);
SetDepthMode(bp);
break;
case SUMEM_BLENDMODE: // Blending Control
{
if (bp.changes & 0xFFFF)
{
PRIM_LOG("blendmode: en=%d, open=%d, colupd=%d, alphaupd=%d, dst=%d, src=%d, sub=%d, mode=%d",
sumem.blendmode.blendenable, sumem.blendmode.logicopenable, sumem.blendmode.colorupdate, sumem.blendmode.alphaupdate,
sumem.blendmode.dstfactor, sumem.blendmode.srcfactor, sumem.blendmode.subtract, sumem.blendmode.logicmode);
// Set LogicOp Blending Mode
if (bp.changes & 2)
{
SETSTAT(stats.logicOpMode, sumem.blendmode.logicopenable != 0 ? sumem.blendmode.logicmode : stats.logicOpMode);
SetLogicOpMode(bp);
}
// Set Dithering Mode
if (bp.changes & 4)
{
SETSTAT(stats.dither, sumem.blendmode.dither);
SetDitherMode(bp);
}
// Set Blending Mode
if (bp.changes & 0xFE1)
{
SETSTAT(stats.srcFactor, sumem.blendmode.srcfactor);
SETSTAT(stats.dstFactor, sumem.blendmode.dstfactor);
SetBlendMode(bp);
}
// Set Color Mask
if (bp.changes & 0x18)
{
SETSTAT(stats.alphaUpdate, sumem.blendmode.alphaupdate);
SETSTAT(stats.colorUpdate, sumem.blendmode.colorupdate);
SetColorMask(bp);
}
}
break;
}
case SUMEM_CONSTANTALPHA: // Set Destination Alpha
{
PRIM_LOG("constalpha: alp=%d, en=%d", sumem.dstalpha.alpha, sumem.dstalpha.enable);
SETSTAT(stats.dstAlphaEnable, sumem.dstalpha.enable);
SETSTAT_UINT(stats.dstAlpha, sumem.dstalpha.alpha);
PixelShaderManager::SetDestAlpha(sumem.dstalpha);
break;
}
// This is called when the game is done drawing the new frame (eg: like in DX: Begin(); Draw(); End();)
case SUMEM_SETDRAWDONE:
switch (bp.newvalue & 0xFF)
{
case 0x02:
g_VideoInitialize.pSetPEFinish(); // may generate interrupt
DEBUG_LOG(VIDEO, "GXSetDrawDone SetPEFinish (value: 0x%02X)", (bp.newvalue & 0xFFFF));
break;
default:
WARN_LOG(VIDEO, "GXSetDrawDone ??? (value 0x%02X)", (bp.newvalue & 0xFFFF));
break;
}
break;
case SUMEM_PE_TOKEN_ID: // Pixel Engine Token ID
g_VideoInitialize.pSetPEToken(static_cast<u16>(bp.newvalue & 0xFFFF), FALSE);
DEBUG_LOG(VIDEO, "SetPEToken 0x%04x", (bp.newvalue & 0xFFFF));
break;
case SUMEM_PE_TOKEN_INT_ID: // Pixel Engine Interrupt Token ID
g_VideoInitialize.pSetPEToken(static_cast<u16>(bp.newvalue & 0xFFFF), TRUE);
DEBUG_LOG(VIDEO, "SetPEToken + INT 0x%04x", (bp.newvalue & 0xFFFF));
break;
// ------------------------
// EFB copy command. This copies a rectangle from the EFB to either RAM in a texture format or to XFB as YUYV.
// It can also optionally clear the EFB while copying from it. To emulate this, we of course copy first and clear afterwards.
case SUMEM_TRIGGER_EFB_COPY: // Copy EFB Region or Render to the XFB or Clear the screen.
{
DVSTARTSUBPROFILE("LoadSUReg:swap");
// The bottom right is within the rectangle
// The values in sumem.copyTexSrcXY and sumem.copyTexSrcWH are updated in case 0x49 and 0x4a in this function
TRectangle rc = {
(int)(sumem.copyTexSrcXY.x),
(int)(sumem.copyTexSrcXY.y),
(int)((sumem.copyTexSrcXY.x + sumem.copyTexSrcWH.x + 1)),
(int)((sumem.copyTexSrcXY.y + sumem.copyTexSrcWH.y + 1))
};
float MValueX = GetRendererTargetScaleX();
float MValueY = GetRendererTargetScaleY();
// Need another rc here to get it to scale.
// Here the bottom right is the out of the rectangle.
TRectangle multirc = {
(int)(sumem.copyTexSrcXY.x * MValueX),
(int)(sumem.copyTexSrcXY.y * MValueY),
(int)((sumem.copyTexSrcXY.x * MValueX + (sumem.copyTexSrcWH.x + 1) * MValueX)),
(int)((sumem.copyTexSrcXY.y * MValueY + (sumem.copyTexSrcWH.y + 1) * MValueY))
};
UPE_Copy PE_copy;
PE_copy.Hex = sumem.triggerEFBCopy;
// Check if we are to copy from the EFB or draw to the XFB
if (PE_copy.copy_to_xfb == 0)
{
if (GetConfig(CONFIG_SHOWEFBREGIONS))
stats.efb_regions.push_back(rc);
CopyEFB(bp, rc, sumem.copyTexDest << 5,
sumem.zcontrol.pixel_format == PIXELFMT_Z24,
PE_copy.intensity_fmt > 0,
((PE_copy.target_pixel_format / 2) + ((PE_copy.target_pixel_format & 1) * 8)),
PE_copy.half_scale > 0);
}
else
{
// the number of lines copied is determined by the y scale * source efb height
const float yScale = sumem.dispcopyyscale / 256.0f;
const float xfbLines = ((sumem.copyTexSrcWH.y + 1.0f) * yScale);
RenderToXFB(bp, multirc, yScale, xfbLines,
Memory_GetPtr(sumem.copyTexDest << 5),
sumem.copyMipMapStrideChannels << 4,
(u32)ceil(xfbLines));
}
// Clear the picture after it's done and submitted, to prepare for the next picture
if (PE_copy.clear)
ClearScreen(bp, multirc);
RestoreRenderState(bp);
break;
}
case SUMEM_LOADTLUT0: // Load a Texture Look Up Table
case SUMEM_LOADTLUT1:
{
DVSTARTSUBPROFILE("LoadSUReg:GXLoadTlut");
u32 tlutTMemAddr = (bp.newvalue & 0x3FF) << 9;
u32 tlutXferCount = (bp.newvalue & 0x1FFC00) >> 5;
u8 *ptr = 0;
// TODO - figure out a cleaner way.
if (GetConfig(CONFIG_ISWII))
ptr = GetPointer(sumem.tlutXferSrc << 5);
else
ptr = GetPointer((sumem.tlutXferSrc & 0xFFFFF) << 5);
if (ptr)
memcpy_gc(texMem + tlutTMemAddr, ptr, tlutXferCount);
else
PanicAlert("Invalid palette pointer %08x %08x %08x", sumem.tlutXferSrc, sumem.tlutXferSrc << 5, (sumem.tlutXferSrc & 0xFFFFF)<< 5);
// TODO(ector) : kill all textures that use this palette
// Not sure if it's a good idea, though. For now, we hash texture palettes
break;
}
case SUMEM_FOGRANGE: // Fog Settings Control
case SUMEM_FOGPARAM0:
case SUMEM_FOGBMAGNITUDE:
case SUMEM_FOGBEXPONENT:
case SUMEM_FOGPARAM3:
if(!GetConfig(CONFIG_DISABLEFOG))
PixelShaderManager::SetFogParamChanged();
break;
case SUMEM_FOGCOLOR: // Fog Color
PixelShaderManager::SetFogColorChanged();
break;
case SUMEM_ALPHACOMPARE: // Compare Alpha Values
PRIM_LOG("alphacmp: ref0=%d, ref1=%d, comp0=%d, comp1=%d, logic=%d", sumem.alphaFunc.ref0,
sumem.alphaFunc.ref1, sumem.alphaFunc.comp0, sumem.alphaFunc.comp1, sumem.alphaFunc.logic);
PixelShaderManager::SetAlpha(sumem.alphaFunc);
break;
case SUMEM_BIAS: // BIAS
PRIM_LOG("ztex bias=0x%x", sumem.ztex1.bias);
PixelShaderManager::SetZTextureBias(sumem.ztex1.bias);
break;
case SUMEM_ZTEX2: // Z Texture type
{
if (bp.changes & 3)
PixelShaderManager::SetZTextureTypeChanged();
#if defined(_DEBUG) || defined(DEBUGFAST)
const char* pzop[] = {"DISABLE", "ADD", "REPLACE", "?"};
const char* pztype[] = {"Z8", "Z16", "Z24", "?"};
PRIM_LOG("ztex op=%s, type=%s", pzop[sumem.ztex2.op], pztype[sumem.ztex2.type]);
#endif
break;
}
// ----------------------------------
// Display Copy Filtering Control
// Fields: Destination, Frame2Field, Gamma, Source
// TODO: We might have to implement the gamma one, some games might need this, if they are too dark to see.
// ----------------------------------
case SUMEM_DISPLAYCOPYFILER:
case SUMEM_DISPLAYCOPYFILER+1:
case SUMEM_DISPLAYCOPYFILER+2:
case SUMEM_DISPLAYCOPYFILER+3:
case SUMEM_COPYFILTER0: //GXSetCopyFilter
case SUMEM_COPYFILTER1:
break;
case SUMEM_FIELDMASK: // Interlacing Control
case SUMEM_FIELDMODE:
SetInterlacingMode(bp);
break;
// ---------------------------------------------------
// Debugging/Profiling info, we don't care about them
// ---------------------------------------------------
case SUMEM_CLOCK0: // Some Clock
case SUMEM_CLOCK1: // Some Clock
case SUMEM_SU_COUNTER: // Pixel or Poly Count
case SUMEM_RAS_COUNTER: // Sound Count of something in the Texture Units
case SUMEM_SETGPMETRIC: // Set the Graphic Processor Metric
break;
// ----------------
// EFB Copy config
// ----------------
case SUMEM_EFB_TL: // EFB Source Rect. Top, Left
case SUMEM_EFB_BR: // EFB Source Rect. Bottom, Right (w, h - 1)
case SUMEM_EFB_ADDR: // EFB Target Address
break;
// --------------
// Clear Config
// --------------
case SUMEM_CLEAR_AR: // Alpha and Red Components
case SUMEM_CLEAR_GB: // Green and Blue Components
case SUMEM_CLEAR_Z: // Z Components (24-bit Zbuffer)
break;
// -------------------------
// Culling Occulsion, we don't support this
// let's hope not many games use bboxes..
// TODO(ector): add something that watches bboxes
// -------------------------
case SUMEM_CLEARBBOX1:
case SUMEM_CLEARBBOX2:
break;
case SUMEM_ZCOMPARE: // Set the Z-Compare
case SUMEM_TEXINVALIDATE: // Used, if game has manual control the Texture Cache, which we don't allow
case SUMEM_MIPMAP_STRIDE: // MipMap Stride Channel
case SUMEM_COPYYSCALE: // Display Copy Y Scale
case SUMEM_IREF: /* 24 RID
21 BC3 - Ind. Tex Stage 3 NTexCoord
18 BI3 - Ind. Tex Stage 3 NTexMap
15 BC2 - Ind. Tex Stage 2 NTexCoord
12 BI2 - Ind. Tex Stage 2 NTexMap
9 BC1 - Ind. Tex Stage 1 NTexCoord
6 BI1 - Ind. Tex Stage 1 NTexMap
3 BC0 - Ind. Tex Stage 0 NTexCoord
0 BI0 - Ind. Tex Stage 0 NTexMap */
break;
case SUMEM_TEV_KSEL: // Texture Environment Swap Mode Table 0
case SUMEM_TEV_KSEL+1:// Texture Environment Swap Mode Table 1
case SUMEM_TEV_KSEL+2:// Texture Environment Swap Mode Table 2
case SUMEM_TEV_KSEL+3:// Texture Environment Swap Mode Table 3
case SUMEM_TEV_KSEL+4:// Texture Environment Swap Mode Table 4
case SUMEM_TEV_KSEL+5:// Texture Environment Swap Mode Table 5
case SUMEM_TEV_KSEL+6:// Texture Environment Swap Mode Table 6
case SUMEM_TEV_KSEL+7:// Texture Environment Swap Mode Table 7
break;
case SUMEM_BP_MASK: // This Register can be used to limit to which bits of BP registers is actually written to. the mask is
// only valid for the next BP command, and will reset itself.
case SUMEM_IND_IMASK: // Index Mask ?
break;
case SUMEM_UNKNOWN: // This is always set to 0xF at boot of any game, so this sounds like a useless reg
if (bp.newvalue != 0x0F)
PanicAlert("Unknown is not 0xF! val = 0x%08x", bp.newvalue);
break;
// Cases added due to: http://code.google.com/p/dolphin-emu/issues/detail?id=360#c90
// Are these related to BBox?
case SUMEM_UNKNOWN1:
case SUMEM_UNKNOWN2:
case SUMEM_UNKNOWN3:
case SUMEM_UNKNOWN4:
break;
// ------------------------------------------------
// On Default, we try to look for other things
// before we give up and say its an unknown opcode
// ------------------------------------------------
default:
switch (bp.address & 0xFC) // Texture sampler filter
{
// -------------------------
// Texture Environment Order
// -------------------------
case SUMEM_TREF:
case SUMEM_TREF+1:
case SUMEM_TREF+2:
case SUMEM_TREF+3:
case SUMEM_TREF+4:
case SUMEM_TREF+5:
case SUMEM_TREF+6:
case SUMEM_TREF+7:
break;
// ----------------------
// Set wrap size
// ----------------------
case SUMEM_SU_SSIZE:
case SUMEM_SU_TSIZE:
case SUMEM_SU_SSIZE+2:
case SUMEM_SU_TSIZE+2:
case SUMEM_SU_SSIZE+4:
case SUMEM_SU_TSIZE+4:
case SUMEM_SU_SSIZE+6:
case SUMEM_SU_TSIZE+6:
case SUMEM_SU_SSIZE+8:
case SUMEM_SU_TSIZE+8:
case SUMEM_SU_SSIZE+10:
case SUMEM_SU_TSIZE+10:
case SUMEM_SU_SSIZE+12:
case SUMEM_SU_TSIZE+12:
case SUMEM_SU_SSIZE+14:
case SUMEM_SU_TSIZE+14:
PixelShaderManager::SetTexCoordChanged((bp.address - SUMEM_SU_SSIZE) >> 1);
break;
// ------------------------
// SUMEM_TX_SETMODE0 - (Texture lookup and filtering mode) LOD/BIAS Clamp, MaxAnsio, LODBIAS, DiagLoad, Min Filter, Mag Filter, Wrap T, S
// SUMEM_TX_SETMODE1 - (LOD Stuff) - Max LOD, Min LOD
// ------------------------
case SUMEM_TX_SETMODE0: // (0x90 for linear)
case SUMEM_TX_SETMODE0+1:
case SUMEM_TX_SETMODE0+2:
case SUMEM_TX_SETMODE0+3:
case SUMEM_TX_SETMODE1:
case SUMEM_TX_SETMODE1+1:
case SUMEM_TX_SETMODE1+2:
case SUMEM_TX_SETMODE1+3:
case SUMEM_TX_SETMODE0_4:
case SUMEM_TX_SETMODE0_4+1:
case SUMEM_TX_SETMODE0_4+2:
case SUMEM_TX_SETMODE0_4+3:
case SUMEM_TX_SETMODE1_4:
case SUMEM_TX_SETMODE1_4+1:
case SUMEM_TX_SETMODE1_4+2:
case SUMEM_TX_SETMODE1_4+3:
SetSamplerState(bp);
break;
// --------------------------------------------
// SUMEM_TX_SETIMAGE0 - Texture width, height, format
// SUMEM_TX_SETIMAGE1 - even LOD address in TMEM - Image Type, Cache Height, Cache Width, TMEM Offset
// SUMEM_TX_SETIMAGE2 - odd LOD address in TMEM - Cache Height, Cache Width, TMEM Offset
// SUMEM_TX_SETIMAGE3 - Address of Texture in main memory
// --------------------------------------------
case SUMEM_TX_SETIMAGE0:
case SUMEM_TX_SETIMAGE0+1:
case SUMEM_TX_SETIMAGE0+2:
case SUMEM_TX_SETIMAGE0+3:
case SUMEM_TX_SETIMAGE0_4:
case SUMEM_TX_SETIMAGE0_4+1:
case SUMEM_TX_SETIMAGE0_4+2:
case SUMEM_TX_SETIMAGE0_4+3:
case SUMEM_TX_SETIMAGE1:
case SUMEM_TX_SETIMAGE1+1:
case SUMEM_TX_SETIMAGE1+2:
case SUMEM_TX_SETIMAGE1+3:
case SUMEM_TX_SETIMAGE1_4:
case SUMEM_TX_SETIMAGE1_4+1:
case SUMEM_TX_SETIMAGE1_4+2:
case SUMEM_TX_SETIMAGE1_4+3:
case SUMEM_TX_SETIMAGE2:
case SUMEM_TX_SETIMAGE2+1:
case SUMEM_TX_SETIMAGE2+2:
case SUMEM_TX_SETIMAGE2+3:
case SUMEM_TX_SETIMAGE2_4:
case SUMEM_TX_SETIMAGE2_4+1:
case SUMEM_TX_SETIMAGE2_4+2:
case SUMEM_TX_SETIMAGE2_4+3:
case SUMEM_TX_SETIMAGE3:
case SUMEM_TX_SETIMAGE3+1:
case SUMEM_TX_SETIMAGE3+2:
case SUMEM_TX_SETIMAGE3+3:
case SUMEM_TX_SETIMAGE3_4:
case SUMEM_TX_SETIMAGE3_4+1:
case SUMEM_TX_SETIMAGE3_4+2:
case SUMEM_TX_SETIMAGE3_4+3:
break;
// -------------------------------
// Set a TLUT
// SUMEM_TX_SETTLUT - Format, TMEM Offset (offset of TLUT from start of TMEM high bank > > 5)
// -------------------------------
case SUMEM_TX_SETTLUT:
case SUMEM_TX_SETTLUT+1:
case SUMEM_TX_SETTLUT+2:
case SUMEM_TX_SETTLUT+3:
case SUMEM_TX_SETLUT_4:
case SUMEM_TX_SETLUT_4+1:
case SUMEM_TX_SETLUT_4+2:
case SUMEM_TX_SETLUT_4+3:
break;
// ---------------------------------------------------
// Set the TEV Color
// ---------------------------------------------------
case SUMEM_TEV_REGISTER_L: // Reg 1
case SUMEM_TEV_REGISTER_H:
case SUMEM_TEV_REGISTER_L+2: // Reg 2
case SUMEM_TEV_REGISTER_H+2:
case SUMEM_TEV_REGISTER_L+4: // Reg 3
case SUMEM_TEV_REGISTER_H+4:
case SUMEM_TEV_REGISTER_L+6: // Reg 4
case SUMEM_TEV_REGISTER_H+6:
{
if (bp.address & 1)
{
// don't compare with changes!
int num = (bp.address >> 1 ) & 0x3;
PixelShaderManager::SetColorChanged(sumem.tevregs[num].high.type, num);
}
break;
}
// ------------------------------------------------
// On Default, we try to look for other things
// before we give up and say its an unknown opcode
// again ...
// ------------------------------------------------
default:
switch (bp.address & 0xF0)
{
// --------------
// Indirect Tev
// --------------
case SUMEM_IND_CMD:
case SUMEM_IND_CMD+1:
case SUMEM_IND_CMD+2:
case SUMEM_IND_CMD+3:
case SUMEM_IND_CMD+4:
case SUMEM_IND_CMD+5:
case SUMEM_IND_CMD+6:
case SUMEM_IND_CMD+7:
case SUMEM_IND_CMD+8:
case SUMEM_IND_CMD+9:
case SUMEM_IND_CMD+10:
case SUMEM_IND_CMD+11:
case SUMEM_IND_CMD+12:
case SUMEM_IND_CMD+13:
case SUMEM_IND_CMD+14:
case SUMEM_IND_CMD+15:
break;
// --------------------------------------------------
// Set Color/Alpha of a Tev
// SUMEM_TEV_COLOR_ENV - Dest, Shift, Clamp, Sub, Bias, Sel A, Sel B, Sel C, Sel D
// SUMEM_TEV_ALPHA_ENV - Dest, Shift, Clamp, Sub, Bias, Sel A, Sel B, Sel C, Sel D, T Swap, R Swap
// --------------------------------------------------
case SUMEM_TEV_COLOR_ENV: // Texture Environment 1
case SUMEM_TEV_ALPHA_ENV:
case SUMEM_TEV_COLOR_ENV+2: // Texture Environment 2
case SUMEM_TEV_ALPHA_ENV+2:
case SUMEM_TEV_COLOR_ENV+4: // Texture Environment 3
case SUMEM_TEV_ALPHA_ENV+4:
case SUMEM_TEV_COLOR_ENV+8: // Texture Environment 4
case SUMEM_TEV_ALPHA_ENV+8:
case SUMEM_TEV_COLOR_ENV+10: // Texture Environment 5
case SUMEM_TEV_ALPHA_ENV+10:
case SUMEM_TEV_COLOR_ENV+12: // Texture Environment 6
case SUMEM_TEV_ALPHA_ENV+12:
case SUMEM_TEV_COLOR_ENV+14: // Texture Environment 7
case SUMEM_TEV_ALPHA_ENV+14:
case SUMEM_TEV_COLOR_ENV+16: // Texture Environment 8
case SUMEM_TEV_ALPHA_ENV+16:
case SUMEM_TEV_COLOR_ENV+18: // Texture Environment 9
case SUMEM_TEV_ALPHA_ENV+18:
case SUMEM_TEV_COLOR_ENV+20: // Texture Environment 10
case SUMEM_TEV_ALPHA_ENV+20:
case SUMEM_TEV_COLOR_ENV+22: // Texture Environment 11
case SUMEM_TEV_ALPHA_ENV+22:
case SUMEM_TEV_COLOR_ENV+24: // Texture Environment 12
case SUMEM_TEV_ALPHA_ENV+24:
case SUMEM_TEV_COLOR_ENV+26: // Texture Environment 13
case SUMEM_TEV_ALPHA_ENV+26:
case SUMEM_TEV_COLOR_ENV+28: // Texture Environment 14
case SUMEM_TEV_ALPHA_ENV+28:
case SUMEM_TEV_COLOR_ENV+30: // Texture Environment 15
case SUMEM_TEV_ALPHA_ENV+30:
case SUMEM_TEV_COLOR_ENV+32: // Texture Environment 16
case SUMEM_TEV_ALPHA_ENV+32:
break;
default:
WARN_LOG(VIDEO, "Unknown Bypass opcode: address = 0x%08x value = 0x%08x", bp.address, bp.newvalue);
break;
}
}
}
}

4
Source/Core/VideoCommon/Src/VertexLoader.cpp
View file

@ -29,7 +29,7 @@
#include "Statistics.h"
#include "VertexLoaderManager.h"
#include "VertexLoader.h"
#include "SUMemory.h"
#include "BPMemory.h"
#include "DataReader.h"
#include "NativeVertexWriter.h"
@ -539,7 +539,7 @@ void VertexLoader::RunVertices(int vtx_attr_group, int primitive, int count)
}
g_nativeVertexFmt = m_NativeFmt;
if (sumem.genMode.cullmode == 3 && primitive < 5)
if (bpmem.genMode.cullmode == 3 && primitive < 5)
{
// if cull mode is none, ignore triangles and quads
DataSkip(count * m_VertexSize);

14
Source/Core/VideoCommon/Src/VertexShaderGen.cpp
View file

@ -20,7 +20,7 @@
#include "Profiler.h"
#include "NativeVertexFormat.h"
#include "SUMemory.h"
#include "BPMemory.h"
#include "VertexShaderGen.h"
// Mash together all the inputs that contribute to the code of a generated vertex shader into
@ -42,8 +42,8 @@ void GetVertexShaderId(VERTEXSHADERUID& vid, u32 components)
}
// fog
vid.values[1] |= (((u32)sumem.fog.c_proj_fsel.fsel & 3) << 30);
vid.values[2] |= (((u32)sumem.fog.c_proj_fsel.fsel >> 2) << 30);
vid.values[1] |= (((u32)bpmem.fog.c_proj_fsel.fsel & 3) << 30);
vid.values[2] |= (((u32)bpmem.fog.c_proj_fsel.fsel >> 2) << 30);
u32* pcurvalue = &vid.values[3];
for (int i = 0; i < xfregs.numTexGens; ++i) {
@ -81,8 +81,8 @@ const char *GenerateVertexShader(u32 components)
text[sizeof(text) - 1] = 0x7C; // canary
DVSTARTPROFILE();
_assert_( sumem.genMode.numtexgens == xfregs.numTexGens);
_assert_( sumem.genMode.numcolchans == xfregs.nNumChans);
_assert_( bpmem.genMode.numtexgens == xfregs.numTexGens);
_assert_( bpmem.genMode.numcolchans == xfregs.nNumChans);
u32 lightMask = 0;
if (xfregs.nNumChans > 0)
@ -432,8 +432,8 @@ const char *GenerateVertexShader(u32 components)
WRITE(p, "o.tex3.w = o.pos.w;\n");
}
// if (sumem.fog.c_proj_fsel.fsel != 0) {
// switch (sumem.fog.c_proj_fsel.fsel) {
// if (bpmem.fog.c_proj_fsel.fsel != 0) {
// switch (bpmem.fog.c_proj_fsel.fsel) {
// case 1: // linear
// break;
// case 4: // exp

2
Source/Core/VideoCommon/Src/VertexShaderManager.cpp
View file

@ -25,7 +25,7 @@
#include "VertexShaderGen.h"
#include "VertexShaderManager.h"
#include "SUMemory.h"
#include "BPMemory.h"
#include "CPMemory.h"
#include "XFMemory.h"
#include "VideoCommon.h"

4
Source/Core/VideoCommon/Src/VideoState.cpp
View file

@ -17,7 +17,7 @@
#include "VideoState.h"
#include "SUMemory.h"
#include "BPMemory.h"
#include "CPMemory.h"
#include "XFMemory.h"
#include "TextureDecoder.h"
@ -26,7 +26,7 @@
static void DoState(PointerWrap &p)
{
// BP Memory
p.Do(sumem);
p.Do(bpmem);
// CP Memory
p.DoArray(arraybases, 16);
p.DoArray(arraystrides, 16);

74
Source/Core/VideoCommon/VideoCommon.vcproj
View file

@ -428,6 +428,14 @@
<Filter
Name="Sections"
>
<File
RelativePath=".\Src\BPMemory.cpp"
>
</File>
<File
RelativePath=".\Src\BPMemory.h"
>
</File>
<File
RelativePath=".\Src\CPMemory.cpp"
>
@ -436,26 +444,6 @@
RelativePath=".\Src\CPMemory.h"
>
</File>
<File
RelativePath=".\Src\SUFunctions.h"
>
</File>
<File
RelativePath=".\Src\SUMemory.cpp"
>
</File>
<File
RelativePath=".\Src\SUMemory.h"
>
</File>
<File
RelativePath=".\Src\SUStructs.cpp"
>
</File>
<File
RelativePath=".\Src\SUStructs.h"
>
</File>
<File
RelativePath=".\Src\XFMemory.cpp"
>
@ -621,26 +609,18 @@
>
</File>
</Filter>
<Filter
Name="Decoding"
<File
RelativePath=".\Src\BPFunctions.h"
>
<File
RelativePath=".\Src\OpcodeDecoding.cpp"
>
</File>
<File
RelativePath=".\Src\OpcodeDecoding.h"
>
</File>
<File
RelativePath=".\Src\TextureDecoder.cpp"
>
</File>
<File
RelativePath=".\Src\TextureDecoder.h"
>
</File>
</Filter>
</File>
<File
RelativePath=".\Src\BPStructs.cpp"
>
</File>
<File
RelativePath=".\Src\BPStructs.h"
>
</File>
<File
RelativePath=".\Src\DataReader.h"
>
@ -677,10 +657,26 @@
RelativePath=".\Src\NativeVertexWriter.h"
>
</File>
<File
RelativePath=".\Src\OpcodeDecoding.cpp"
>
</File>
<File
RelativePath=".\Src\OpcodeDecoding.h"
>
</File>
<File
RelativePath=".\Src\SConscript"
>
</File>
<File
RelativePath=".\Src\TextureDecoder.cpp"
>
</File>
<File
RelativePath=".\Src\TextureDecoder.h"
>
</File>
<File
RelativePath=".\Src\VertexLoaderManager.cpp"
>

2
Source/Plugins/Plugin_VideoDX9/Plugin_VideoDX9.vcproj
View file

@ -1180,7 +1180,7 @@
Name="Decoder"
>
<File
RelativePath=".\Src\SUFunctions.cpp"
RelativePath=".\Src\BPFunctions.cpp"
>
</File>
</Filter>

88
Source/Plugins/Plugin_VideoDX9/Src/SUFunctions.cpp → Source/Plugins/Plugin_VideoDX9/Src/BPFunctions.cpp
View file

@ -15,7 +15,7 @@
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#include "SUFunctions.h"
#include "BPFunctions.h"
#include "D3DBase.h"
#include "Config.h"
#include "Common.h"
@ -92,7 +92,7 @@ static const D3DTEXTUREADDRESS d3dClamps[4] =
D3DTADDRESS_WRAP //reserved
};
namespace SUFunctions
namespace BPFunctions
{
void FlushPipeline()
@ -100,12 +100,12 @@ void FlushPipeline()
VertexManager::Flush();
}
void SetGenerationMode(const BPCommand &bp)
void SetGenerationMode(const Bypass &bp)
{
// dev->SetRenderState(D3DRS_CULLMODE, d3dCullModes[sumem.genMode.cullmode]);
Renderer::SetRenderState(D3DRS_CULLMODE, d3dCullModes[sumem.genMode.cullmode]);
// dev->SetRenderState(D3DRS_CULLMODE, d3dCullModes[bpmem.genMode.cullmode]);
Renderer::SetRenderState(D3DRS_CULLMODE, d3dCullModes[bpmem.genMode.cullmode]);
if (sumem.genMode.cullmode == 3)
if (bpmem.genMode.cullmode == 3)
{
// dev->SetRenderState(D3DRS_COLORWRITEENABLE, 0);
Renderer::SetRenderState(D3DRS_COLORWRITEENABLE, 0);
@ -113,9 +113,9 @@ void SetGenerationMode(const BPCommand &bp)
else
{
DWORD write = 0;
if (sumem.blendmode.alphaupdate)
if (bpmem.blendmode.alphaupdate)
write = D3DCOLORWRITEENABLE_ALPHA;
if (sumem.blendmode.colorupdate)
if (bpmem.blendmode.colorupdate)
write |= D3DCOLORWRITEENABLE_RED | D3DCOLORWRITEENABLE_GREEN | D3DCOLORWRITEENABLE_BLUE;
// dev->SetRenderState(D3DRS_COLORWRITEENABLE, write);
@ -123,27 +123,27 @@ void SetGenerationMode(const BPCommand &bp)
}
}
void SetScissor(const BPCommand &bp)
void SetScissor(const Bypass &bp)
{
Renderer::SetScissorRect();
}
void SetLineWidth(const BPCommand &bp)
void SetLineWidth(const Bypass &bp)
{
// We can't change line width in D3D unless we use ID3DXLine
float psize = float(sumem.lineptwidth.pointsize) * 6.0f;
float psize = float(bpmem.lineptwidth.pointsize) * 6.0f;
Renderer::SetRenderState(D3DRS_POINTSIZE, *((DWORD*)&psize));
}
void SetDepthMode(const BPCommand &bp)
void SetDepthMode(const Bypass &bp)
{
if (sumem.zmode.testenable)
if (bpmem.zmode.testenable)
{
// dev->SetRenderState(D3DRS_ZENABLE, TRUE);
// dev->SetRenderState(D3DRS_ZWRITEENABLE, sumem.zmode.updateenable);
// dev->SetRenderState(D3DRS_ZFUNC,d3dCmpFuncs[sumem.zmode.func]);
// dev->SetRenderState(D3DRS_ZWRITEENABLE, bpmem.zmode.updateenable);
// dev->SetRenderState(D3DRS_ZFUNC,d3dCmpFuncs[bpmem.zmode.func]);
Renderer::SetRenderState(D3DRS_ZENABLE, TRUE);
Renderer::SetRenderState(D3DRS_ZWRITEENABLE, sumem.zmode.updateenable);
Renderer::SetRenderState(D3DRS_ZFUNC, d3dCmpFuncs[sumem.zmode.func]);
Renderer::SetRenderState(D3DRS_ZWRITEENABLE, bpmem.zmode.updateenable);
Renderer::SetRenderState(D3DRS_ZFUNC, d3dCmpFuncs[bpmem.zmode.func]);
}
else
{
@ -155,23 +155,23 @@ void SetDepthMode(const BPCommand &bp)
Renderer::SetRenderState(D3DRS_ZWRITEENABLE, FALSE);
}
//if (!sumem.zmode.updateenable)
//if (!bpmem.zmode.updateenable)
// Renderer::SetRenderMode(Renderer::RM_Normal);
}
void SetBlendMode(const BPCommand &bp)
void SetBlendMode(const Bypass &bp)
{
if (bp.changes & 1)
Renderer::SetRenderState(D3DRS_ALPHABLENDENABLE, sumem.blendmode.blendenable);
Renderer::SetRenderState(D3DRS_ALPHABLENDENABLE, bpmem.blendmode.blendenable);
D3DBLEND src = d3dSrcFactors[sumem.blendmode.srcfactor];
D3DBLEND dst = d3dDestFactors[sumem.blendmode.dstfactor];
D3DBLEND src = d3dSrcFactors[bpmem.blendmode.srcfactor];
D3DBLEND dst = d3dDestFactors[bpmem.blendmode.dstfactor];
if (bp.changes & 0x700)
Renderer::SetRenderState(D3DRS_SRCBLEND, src);
if (bp.changes & 0xE0) {
if (!sumem.blendmode.subtract)
if (!bpmem.blendmode.subtract)
{
Renderer::SetRenderState(D3DRS_DESTBLEND, dst);
}
@ -182,7 +182,7 @@ void SetBlendMode(const BPCommand &bp)
}
if (bp.changes & 0x800)
{
if (sumem.blendmode.subtract)
if (bpmem.blendmode.subtract)
{
Renderer::SetRenderState(D3DRS_SRCBLEND, D3DBLEND_ONE);
Renderer::SetRenderState(D3DRS_DESTBLEND, D3DBLEND_ONE);
@ -193,23 +193,23 @@ void SetBlendMode(const BPCommand &bp)
Renderer::SetRenderState(D3DRS_DESTBLEND, dst);
}
Renderer::SetRenderState(D3DRS_BLENDOP, sumem.blendmode.subtract ? D3DBLENDOP_SUBTRACT : D3DBLENDOP_ADD);
Renderer::SetRenderState(D3DRS_BLENDOP, bpmem.blendmode.subtract ? D3DBLENDOP_SUBTRACT : D3DBLENDOP_ADD);
}
}
void SetDitherMode(const BPCommand &bp)
void SetDitherMode(const Bypass &bp)
{
Renderer::SetRenderState(D3DRS_DITHERENABLE,sumem.blendmode.dither);
Renderer::SetRenderState(D3DRS_DITHERENABLE,bpmem.blendmode.dither);
}
void SetLogicOpMode(const BPCommand &bp)
void SetLogicOpMode(const Bypass &bp)
{
// Logic op blending. D3D can't do this but can fake some modes.
}
void SetColorMask(const BPCommand &bp)
void SetColorMask(const Bypass &bp)
{
DWORD write = 0;
if (sumem.blendmode.alphaupdate)
if (bpmem.blendmode.alphaupdate)
write = D3DCOLORWRITEENABLE_ALPHA;
if (sumem.blendmode.colorupdate)
if (bpmem.blendmode.colorupdate)
write |= D3DCOLORWRITEENABLE_RED | D3DCOLORWRITEENABLE_GREEN | D3DCOLORWRITEENABLE_BLUE;
Renderer::SetRenderState(D3DRS_COLORWRITEENABLE, write);
@ -222,19 +222,19 @@ float GetRendererTargetScaleY()
{
return Renderer::GetYScale();
}
void CopyEFB(const BPCommand &bp, const TRectangle &rc, const u32 &address, const bool &fromZBuffer, const bool &isIntensityFmt, const u32 &copyfmt, const bool &scaleByHalf)
void CopyEFB(const Bypass &bp, const TRectangle &rc, const u32 &address, const bool &fromZBuffer, const bool &isIntensityFmt, const u32 &copyfmt, const bool &scaleByHalf)
{
RECT rec = { rc.left, rc.top, rc.right, rc.bottom };
TextureCache::CopyEFBToRenderTarget(sumem.copyTexDest<<5, &rec);
TextureCache::CopyEFBToRenderTarget(bpmem.copyTexDest<<5, &rec);
}
void RenderToXFB(const BPCommand &bp, const TRectangle &multirc, const float &yScale, const float &xfbLines, u8* pXFB, const u32 &dstWidth, const u32 &dstHeight)
void RenderToXFB(const Bypass &bp, const TRectangle &multirc, const float &yScale, const float &xfbLines, u8* pXFB, const u32 &dstWidth, const u32 &dstHeight)
{
Renderer::SwapBuffers();
PRIM_LOG("Renderer::SwapBuffers()");
g_VideoInitialize.pCopiedToXFB();
}
void ClearScreen(const BPCommand &bp, const TRectangle &multirc)
void ClearScreen(const Bypass &bp, const TRectangle &multirc)
{
// it seems that the GC is able to alpha blend on color-fill
// we cant do that so if alpha is != 255 we skip it
@ -247,17 +247,17 @@ void ClearScreen(const BPCommand &bp, const TRectangle &multirc)
D3DCOLOR col = 0;
float clearZ = 0;
if (sumem.blendmode.colorupdate || sumem.blendmode.alphaupdate)
if (bpmem.blendmode.colorupdate || bpmem.blendmode.alphaupdate)
{
if (sumem.blendmode.colorupdate || sumem.blendmode.alphaupdate)
col = (sumem.clearcolorAR << 16) | sumem.clearcolorGB;
if (bpmem.blendmode.colorupdate || bpmem.blendmode.alphaupdate)
col = (bpmem.clearcolorAR << 16) | bpmem.clearcolorGB;
// clearflags |= D3DCLEAR_TARGET; set to break animal crossing :p
}
// clear z-buffer
if (sumem.zmode.updateenable)
if (bpmem.zmode.updateenable)
{
clearZ = (float)(sumem.clearZValue & 0xFFFFFF) / float(0xFFFFFF);
clearZ = (float)(bpmem.clearZValue & 0xFFFFFF) / float(0xFFFFFF);
if (clearZ > 1.0f) clearZ = 1.0f;
if (clearZ < 0.0f) clearZ = 0.0f;
clearflags |= D3DCLEAR_ZBUFFER | D3DCLEAR_STENCIL;
@ -266,7 +266,7 @@ void ClearScreen(const BPCommand &bp, const TRectangle &multirc)
D3D::dev->Clear(0, NULL, clearflags, col, clearZ, 0);
}
void RestoreRenderState(const BPCommand &bp)
void RestoreRenderState(const Bypass &bp)
{
//Renderer::SetRenderMode(Renderer::RM_Normal);
}
@ -290,9 +290,9 @@ u8 *GetPointer(const u32 &address)
{
return g_VideoInitialize.pGetMemoryPointer(address);
}
void SetSamplerState(const BPCommand &bp)
void SetSamplerState(const Bypass &bp)
{
FourTexUnits &tex = sumem.tex[(bp.address & 0xE0) == 0xA0];
FourTexUnits &tex = bpmem.tex[(bp.address & 0xE0) == 0xA0];
int stage = (bp.address & 3);//(addr>>4)&2;
TexMode0 &tm0 = tex.texMode0[stage];
@ -329,7 +329,7 @@ void SetSamplerState(const BPCommand &bp)
//sprintf(temp,"lod %f",tm0.lod_bias/4.0f);
//g_VideoInitialize.pLog(temp);
}
void SetInterlacingMode(const BPCommand &bp)
void SetInterlacingMode(const Bypass &bp)
{
// TODO
}

2
Source/Plugins/Plugin_VideoDX9/Src/PixelShaderCache.cpp
View file

@ -25,7 +25,7 @@
#include "PixelShaderManager.h"
#include "PixelShaderCache.h"
#include "VertexLoader.h"
#include "SUMemory.h"
#include "BPMemory.h"
#include "XFMemory.h"
#include <Cg/cg.h>

20
Source/Plugins/Plugin_VideoDX9/Src/Render.cpp
View file

@ -26,7 +26,7 @@
#include "VertexManager.h"
#include "Render.h"
#include "OpcodeDecoding.h"
#include "SUStructs.h"
#include "BPStructs.h"
#include "XFStructs.h"
#include "D3DPostprocess.h"
#include "D3DUtil.h"
@ -386,7 +386,7 @@ void Renderer::SwapBuffers()
D3D::dev->SetRenderState(D3DRS_SCISSORTESTENABLE, false);
D3D::dev->Clear(0, 0, D3DCLEAR_TARGET, 0x101010, 0, 0);
u32 clearColor = (sumem.clearcolorAR << 16) | sumem.clearcolorGB;
u32 clearColor = (bpmem.clearcolorAR << 16) | bpmem.clearcolorGB;
// clearColor |= 0x003F003F;
// D3D::BeginFrame(true,clearColor,1.0f);
D3D::BeginFrame(false, clearColor, 1.0f);
@ -441,13 +441,13 @@ void Renderer::SetViewport(float* _Viewport)
void Renderer::SetScissorRect()
{
int xoff = sumem.scissorOffset.x * 2 - 342;
int yoff = sumem.scissorOffset.y * 2 - 342;
int xoff = bpmem.scissorOffset.x * 2 - 342;
int yoff = bpmem.scissorOffset.y * 2 - 342;
RECT rc;
rc.left = (int)((float)sumem.scissorTL.x - xoff - 342);
rc.top = (int)((float)sumem.scissorTL.y - yoff - 342);
rc.right = (int)((float)sumem.scissorBR.x - xoff - 341);
rc.bottom = (int)((float)sumem.scissorBR.y - yoff - 341);
rc.left = (int)((float)bpmem.scissorTL.x - xoff - 342);
rc.top = (int)((float)bpmem.scissorTL.y - yoff - 342);
rc.right = (int)((float)bpmem.scissorBR.x - xoff - 341);
rc.bottom = (int)((float)bpmem.scissorBR.y - yoff - 341);
rc.left = (int)(rc.left * xScale);
rc.top = (int)(rc.top * yScale);
@ -588,8 +588,8 @@ void UpdateViewport()
// Keep aspect ratio at 4:3
// rawViewport[0] = 320, rawViewport[1] = -240
int scissorXOff = sumem.scissorOffset.x * 2 - 342;
int scissorYOff = sumem.scissorOffset.y * 2 - 342;
int scissorXOff = bpmem.scissorOffset.x * 2 - 342;
int scissorYOff = bpmem.scissorOffset.y * 2 - 342;
float fourThree = 4.0f / 3.0f;
// These were commented out to fix "unreferenced local variable" compiler warnings.
// float wAdj, hAdj;

2
Source/Plugins/Plugin_VideoDX9/Src/TextureCache.cpp
View file

@ -213,7 +213,7 @@ TextureCache::TCacheEntry *TextureCache::Load(int stage, u32 address, int width,
entry.w = width;
entry.h = height;
entry.fmt = format;
entry.mode = sumem.tex[stage > 3].texMode0[stage & 3];
entry.mode = bpmem.tex[stage > 3].texMode0[stage & 3];
if (g_Config.bDumpTextures)
{ // dump texture to file

2
Source/Plugins/Plugin_VideoDX9/Src/TextureCache.h
View file

@ -22,7 +22,7 @@
#include <map>
#include "D3DBase.h"
#include "SUMemory.h"
#include "BPMemory.h"
class TextureCache
{

18
Source/Plugins/Plugin_VideoDX9/Src/VertexManager.cpp
View file

@ -33,7 +33,7 @@
#include "NativeVertexWriter.h"
#include "TextureCache.h"
#include "SUStructs.h"
#include "BPStructs.h"
#include "XFStructs.h"
using namespace D3D;
@ -185,15 +185,15 @@ void Flush()
g_nativeVertexFmt->SetupVertexPointers();
u32 usedtextures = 0;
for (u32 i = 0; i < (u32)sumem.genMode.numtevstages + 1; ++i) {
if (sumem.tevorders[i/2].getEnable(i & 1))
usedtextures |= 1 << sumem.tevorders[i/2].getTexMap(i & 1);
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; ++i) {
if (bpmem.tevorders[i/2].getEnable(i & 1))
usedtextures |= 1 << bpmem.tevorders[i/2].getTexMap(i & 1);
}
if (sumem.genMode.numindstages > 0) {
for (u32 i = 0; i < (u32)sumem.genMode.numtevstages + 1; ++i) {
if (sumem.tevind[i].IsActive() && sumem.tevind[i].bt < sumem.genMode.numindstages) {
usedtextures |= 1 << sumem.tevindref.getTexMap(sumem.tevind[i].bt);
if (bpmem.genMode.numindstages > 0) {
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; ++i) {
if (bpmem.tevind[i].IsActive() && bpmem.tevind[i].bt < bpmem.genMode.numindstages) {
usedtextures |= 1 << bpmem.tevindref.getTexMap(bpmem.tevind[i].bt);
}
}
}
@ -202,7 +202,7 @@ void Flush()
for (int i = 0; i < 8; i++)
{
if (usedtextures & (1 << i)) {
FourTexUnits &tex = sumem.tex[i >> 2];
FourTexUnits &tex = bpmem.tex[i >> 2];
TextureCache::TCacheEntry* tentry = TextureCache::Load(i,
(tex.texImage3[i&3].image_base/* & 0x1FFFFF*/) << 5,
tex.texImage0[i&3].width+1, tex.texImage0[i&3].height+1,

2
Source/Plugins/Plugin_VideoDX9/Src/VertexShaderCache.cpp
View file

@ -25,7 +25,7 @@
#include "Config.h"
#include "VertexShaderCache.h"
#include "VertexLoader.h"
#include "SUMemory.h"
#include "BPMemory.h"
#include "XFMemory.h"
#include <Cg/cg.h>

4
Source/Plugins/Plugin_VideoDX9/Src/main.cpp
View file

@ -28,7 +28,7 @@
#include "Fifo.h"
#include "OpcodeDecoding.h"
#include "TextureCache.h"
#include "SUStructs.h"
#include "BPStructs.h"
#include "VertexManager.h"
#include "VertexLoaderManager.h"
#include "VertexShaderManager.h"
@ -223,7 +223,7 @@ void Video_Prepare(void)
TextureCache::Init();
SUInit();
BPInit();
VertexManager::Init();
Fifo_Init();
VertexLoaderManager::Init();

4
Source/Plugins/Plugin_VideoOGL/Plugin_VideoOGL.vcproj
View file

@ -717,11 +717,11 @@
Name="Decoder"
>
<File
RelativePath=".\Src\NativeVertexFormat.cpp"
RelativePath=".\Src\BPFunctions.cpp"
>
</File>
<File
RelativePath=".\Src\SUFunctions.cpp"
RelativePath=".\Src\NativeVertexFormat.cpp"
>
</File>
<File

70
Source/Plugins/Plugin_VideoOGL/Src/SUFunctions.cpp → Source/Plugins/Plugin_VideoOGL/Src/BPFunctions.cpp
View file

@ -15,7 +15,7 @@
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#include "SUFunctions.h"
#include "BPFunctions.h"
#include "Globals.h"
#include "Profiler.h"
#include "Config.h"
@ -27,7 +27,7 @@
#include "XFB.h"
#include "main.h"
namespace SUFunctions
namespace BPFunctions
{
// ----------------------------------------------
// State translation lookup tables
@ -47,40 +47,40 @@ void FlushPipeline()
{
VertexManager::Flush();
}
void SetGenerationMode(const BPCommand &bp)
void SetGenerationMode(const Bypass &bp)
{
// none, ccw, cw, ccw
if (sumem.genMode.cullmode > 0)
if (bpmem.genMode.cullmode > 0)
{
glEnable(GL_CULL_FACE);
glFrontFace(sumem.genMode.cullmode == 2 ? GL_CCW : GL_CW);
glFrontFace(bpmem.genMode.cullmode == 2 ? GL_CCW : GL_CW);
}
else
glDisable(GL_CULL_FACE);
}
void SetScissor(const BPCommand &bp)
void SetScissor(const Bypass &bp)
{
if (!Renderer::SetScissorRect())
if (bp.address == SUMEM_SCISSORBR)
if (bp.address == BPMEM_SCISSORBR)
ERROR_LOG(VIDEO, "bad scissor!");
}
void SetLineWidth(const BPCommand &bp)
void SetLineWidth(const Bypass &bp)
{
float fratio = xfregs.rawViewport[0] != 0 ? ((float)Renderer::GetTargetWidth() / EFB_WIDTH) : 1.0f;
if (sumem.lineptwidth.linesize > 0)
glLineWidth((float)sumem.lineptwidth.linesize * fratio / 6.0f); // scale by ratio of widths
if (sumem.lineptwidth.pointsize > 0)
glPointSize((float)sumem.lineptwidth.pointsize * fratio / 6.0f);
if (bpmem.lineptwidth.linesize > 0)
glLineWidth((float)bpmem.lineptwidth.linesize * fratio / 6.0f); // scale by ratio of widths
if (bpmem.lineptwidth.pointsize > 0)
glPointSize((float)bpmem.lineptwidth.pointsize * fratio / 6.0f);
}
void SetDepthMode(const BPCommand &bp)
void SetDepthMode(const Bypass &bp)
{
if (sumem.zmode.testenable)
if (bpmem.zmode.testenable)
{
glEnable(GL_DEPTH_TEST);
glDepthMask(sumem.zmode.updateenable ? GL_TRUE : GL_FALSE);
glDepthFunc(glCmpFuncs[sumem.zmode.func]);
glDepthMask(bpmem.zmode.updateenable ? GL_TRUE : GL_FALSE);
glDepthFunc(glCmpFuncs[bpmem.zmode.func]);
}
else
{
@ -89,28 +89,28 @@ void SetDepthMode(const BPCommand &bp)
glDepthMask(GL_FALSE);
}
}
void SetBlendMode(const BPCommand &bp)
void SetBlendMode(const Bypass &bp)
{
Renderer::SetBlendMode(false);
}
void SetDitherMode(const BPCommand &bp)
void SetDitherMode(const Bypass &bp)
{
if (sumem.blendmode.dither)
if (bpmem.blendmode.dither)
glEnable(GL_DITHER);
else
glDisable(GL_DITHER);
}
void SetLogicOpMode(const BPCommand &bp)
void SetLogicOpMode(const Bypass &bp)
{
if (sumem.blendmode.logicopenable)
if (bpmem.blendmode.logicopenable)
{
glEnable(GL_COLOR_LOGIC_OP);
glLogicOp(glLogicOpCodes[sumem.blendmode.logicmode]);
glLogicOp(glLogicOpCodes[bpmem.blendmode.logicmode]);
}
else
glDisable(GL_COLOR_LOGIC_OP);
}
void SetColorMask(const BPCommand &bp)
void SetColorMask(const Bypass &bp)
{
Renderer::SetColorMask();
}
@ -122,9 +122,9 @@ float GetRendererTargetScaleY()
{
return Renderer::GetTargetScaleY();
}
void CopyEFB(const BPCommand &bp, const TRectangle &rc, const u32 &address, const bool &fromZBuffer, const bool &isIntensityFmt, const u32 &copyfmt, const bool &scaleByHalf)
void CopyEFB(const Bypass &bp, const TRectangle &rc, const u32 &address, const bool &fromZBuffer, const bool &isIntensityFmt, const u32 &copyfmt, const bool &scaleByHalf)
{
// sumem.zcontrol.pixel_format to PIXELFMT_Z24 is when the game wants to copy from ZBuffer (Zbuffer uses 24-bit Format)
// bpmem.zcontrol.pixel_format to PIXELFMT_Z24 is when the game wants to copy from ZBuffer (Zbuffer uses 24-bit Format)
if (!g_Config.bEFBCopyDisable)
if (g_Config.bCopyEFBToRAM) // To RAM
TextureConverter::EncodeToRam(address, fromZBuffer, isIntensityFmt, copyfmt, scaleByHalf, rc);
@ -132,7 +132,7 @@ void CopyEFB(const BPCommand &bp, const TRectangle &rc, const u32 &address, cons
TextureMngr::CopyRenderTargetToTexture(address, fromZBuffer, isIntensityFmt, copyfmt, scaleByHalf, rc);
}
void RenderToXFB(const BPCommand &bp, const TRectangle &multirc, const float &yScale, const float &xfbLines, u8* pXFB, const u32 &dstWidth, const u32 &dstHeight)
void RenderToXFB(const Bypass &bp, const TRectangle &multirc, const float &yScale, const float &xfbLines, u8* pXFB, const u32 &dstWidth, const u32 &dstHeight)
{
// EFB to XFB
if (g_Config.bUseXFB)
@ -155,7 +155,7 @@ void RenderToXFB(const BPCommand &bp, const TRectangle &multirc, const float &yS
}
g_VideoInitialize.pCopiedToXFB();
}
void ClearScreen(const BPCommand &bp, const TRectangle &multirc)
void ClearScreen(const Bypass &bp, const TRectangle &multirc)
{
// Update the view port for clearing the picture
glViewport(0, 0, Renderer::GetTargetWidth(), Renderer::GetTargetHeight());
@ -169,21 +169,21 @@ void ClearScreen(const BPCommand &bp, const TRectangle &multirc)
// Since clear operations use the source rectangle, we have to do
// regular renders (glClear clears the entire buffer)
if (sumem.blendmode.colorupdate || sumem.blendmode.alphaupdate || sumem.zmode.updateenable)
if (bpmem.blendmode.colorupdate || bpmem.blendmode.alphaupdate || bpmem.zmode.updateenable)
{
GLbitfield bits = 0;
if (sumem.blendmode.colorupdate || sumem.blendmode.alphaupdate)
if (bpmem.blendmode.colorupdate || bpmem.blendmode.alphaupdate)
{
u32 clearColor = (sumem.clearcolorAR << 16) | sumem.clearcolorGB;
u32 clearColor = (bpmem.clearcolorAR << 16) | bpmem.clearcolorGB;
glClearColor(((clearColor>>16) & 0xff)*(1/255.0f),
((clearColor>>8 ) & 0xff)*(1/255.0f),
((clearColor>>0 ) & 0xff)*(1/255.0f),
((clearColor>>24) & 0xff)*(1/255.0f));
bits |= GL_COLOR_BUFFER_BIT;
}
if (sumem.zmode.updateenable)
if (bpmem.zmode.updateenable)
{
glClearDepth((float)(sumem.clearZValue & 0xFFFFFF) / float(0xFFFFFF));
glClearDepth((float)(bpmem.clearZValue & 0xFFFFFF) / float(0xFFFFFF));
bits |= GL_DEPTH_BUFFER_BIT;
}
glDrawBuffer(GL_COLOR_ATTACHMENT0_EXT);
@ -191,7 +191,7 @@ void ClearScreen(const BPCommand &bp, const TRectangle &multirc)
}
}
void RestoreRenderState(const BPCommand &bp)
void RestoreRenderState(const Bypass &bp)
{
Renderer::RestoreGLState();
}
@ -215,11 +215,11 @@ u8 *GetPointer(const u32 &address)
{
return g_VideoInitialize.pGetMemoryPointer(address);
}
void SetSamplerState(const BPCommand &bp)
void SetSamplerState(const Bypass &bp)
{
// TODO
}
void SetInterlacingMode(const BPCommand &bp)
void SetInterlacingMode(const Bypass &bp)
{
// TODO
}

2
Source/Plugins/Plugin_VideoOGL/Src/PixelShaderCache.h
View file

@ -21,7 +21,7 @@
#include <map>
#include <string>
#include "SUMemory.h"
#include "BPMemory.h"
#include "PixelShaderGen.h"
struct FRAGMENTSHADER

46
Source/Plugins/Plugin_VideoOGL/Src/Render.cpp
View file

@ -38,7 +38,7 @@
#include "ImageWrite.h"
#include "Render.h"
#include "OpcodeDecoding.h"
#include "SUStructs.h"
#include "BPStructs.h"
#include "TextureMngr.h"
#include "rasterfont.h"
#include "VertexShaderGen.h"
@ -606,9 +606,9 @@ void Renderer::ResetGLState()
void Renderer::RestoreGLState()
{
// Gets us back into a more game-like state.
if (sumem.genMode.cullmode > 0) glEnable(GL_CULL_FACE);
if (sumem.zmode.testenable) glEnable(GL_DEPTH_TEST);
if (sumem.zmode.updateenable) glDepthMask(GL_TRUE);
if (bpmem.genMode.cullmode > 0) glEnable(GL_CULL_FACE);
if (bpmem.zmode.testenable) glEnable(GL_DEPTH_TEST);
if (bpmem.zmode.updateenable) glDepthMask(GL_TRUE);
glEnable(GL_SCISSOR_TEST);
SetScissorRect();
@ -621,11 +621,11 @@ void Renderer::RestoreGLState()
void Renderer::SetColorMask()
{
if (sumem.blendmode.alphaupdate && sumem.blendmode.colorupdate)
if (bpmem.blendmode.alphaupdate && bpmem.blendmode.colorupdate)
glColorMask(GL_TRUE, GL_TRUE, GL_TRUE, GL_TRUE);
else if (sumem.blendmode.alphaupdate)
else if (bpmem.blendmode.alphaupdate)
glColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_TRUE);
else if (sumem.blendmode.colorupdate)
else if (bpmem.blendmode.colorupdate)
glColorMask(GL_TRUE, GL_TRUE, GL_TRUE, GL_FALSE);
else
glColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_FALSE);
@ -639,14 +639,14 @@ void Renderer::SetBlendMode(bool forceUpdate)
// 3-5 - srcRGB function
// 6-8 - dstRGB function
u32 newval = sumem.blendmode.subtract << 2;
u32 newval = bpmem.blendmode.subtract << 2;
if (sumem.blendmode.subtract) {
if (bpmem.blendmode.subtract) {
newval |= 0x0049; // enable blending src 1 dst 1
} else if (sumem.blendmode.blendenable) {
} else if (bpmem.blendmode.blendenable) {
newval |= 1; // enable blending
newval |= sumem.blendmode.srcfactor << 3;
newval |= sumem.blendmode.dstfactor << 6;
newval |= bpmem.blendmode.srcfactor << 3;
newval |= bpmem.blendmode.dstfactor << 6;
}
u32 changes = forceUpdate ? 0xFFFFFFFF : newval ^ s_blendMode;
@ -735,30 +735,30 @@ GLuint Renderer::ResolveAndGetDepthTarget(const TRectangle &source_rect)
// Call browser: OpcodeDecoding.cpp ExecuteDisplayList > Decode() > LoadBPReg()
// case 0x52 > SetScissorRect()
// ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
// sumem.scissorTL.x, y = 342x342
// sumem.scissorBR.x, y = 981x821
// bpmem.scissorTL.x, y = 342x342
// bpmem.scissorBR.x, y = 981x821
// Renderer::GetTargetHeight() = the fixed ini file setting
// donkopunchstania - it appears scissorBR is the bottom right pixel inside the scissor box
// therefore the width and height are (scissorBR + 1) - scissorTL
bool Renderer::SetScissorRect()
{
int xoff = sumem.scissorOffset.x * 2 - 342;
int yoff = sumem.scissorOffset.y * 2 - 342;
int xoff = bpmem.scissorOffset.x * 2 - 342;
int yoff = bpmem.scissorOffset.y * 2 - 342;
float MValueX = GetTargetScaleX();
float MValueY = GetTargetScaleY();
float rc_left = (float)sumem.scissorTL.x - xoff - 342; // left = 0
float rc_left = (float)bpmem.scissorTL.x - xoff - 342; // left = 0
rc_left *= MValueX;
if (rc_left < 0) rc_left = 0;
float rc_top = (float)sumem.scissorTL.y - yoff - 342; // right = 0
float rc_top = (float)bpmem.scissorTL.y - yoff - 342; // right = 0
rc_top *= MValueY;
if (rc_top < 0) rc_top = 0;
float rc_right = (float)sumem.scissorBR.x - xoff - 341; // right = 640
float rc_right = (float)bpmem.scissorBR.x - xoff - 341; // right = 640
rc_right *= MValueX;
if (rc_right > EFB_WIDTH * MValueX) rc_right = EFB_WIDTH * MValueX;
float rc_bottom = (float)sumem.scissorBR.y - yoff - 341; // bottom = 480
float rc_bottom = (float)bpmem.scissorBR.y - yoff - 341; // bottom = 480
rc_bottom *= MValueY;
if (rc_bottom > EFB_HEIGHT * MValueY) rc_bottom = EFB_HEIGHT * MValueY;
@ -1384,7 +1384,7 @@ void Renderer::SetScreenshot(const char *filename)
bool Renderer::SaveRenderTarget(const char *filename, int W, int H, int YOffset)
{
// The height seemed to often be one less than the setting (but sometimes not),
// perhaps the source is the (sumem.copyTexSrcWH.y + 1) in SUStructs.cpp that I'm guessing
// perhaps the source is the (bpmem.copyTexSrcWH.y + 1) in BPStructs.cpp that I'm guessing
// is there because of how some GL function works. But the buffer we are reading from here
// seems to have the necessary pixels for a complete height so we use the complete height
// from the settings.
@ -1495,8 +1495,8 @@ void UpdateViewport()
(rawViewport[5] - rawViewport[2]) / 16777215.0f, rawViewport[5] / 16777215.0f);*/
// --------
int scissorXOff = sumem.scissorOffset.x * 2 - 342;
int scissorYOff = sumem.scissorOffset.y * 2 - 342;
int scissorXOff = bpmem.scissorOffset.x * 2 - 342;
int scissorYOff = bpmem.scissorOffset.y * 2 - 342;
// -------------------------------------
float MValueX = Renderer::GetTargetScaleX();

2
Source/Plugins/Plugin_VideoOGL/Src/SConscript
View file

@ -10,7 +10,7 @@ import platform
name = "Plugin_VideoOGL"
files = [
'SUFunctions.cpp',
'BPFunctions.cpp',
'Config.cpp',
'rasterfont.cpp',
'Render.cpp',

2
Source/Plugins/Plugin_VideoOGL/Src/TextureConversionShader.cpp
View file

@ -20,7 +20,7 @@
#include "TextureDecoder.h"
#include "PixelShaderManager.h"
#include "PixelShaderGen.h"
#include "SUMemory.h"
#include "BPMemory.h"
#include <stdio.h>
#include <math.h>

4
Source/Plugins/Plugin_VideoOGL/Src/TextureMngr.cpp
View file

@ -42,7 +42,7 @@
#include "Render.h"
#include "MemoryUtil.h"
#include "SUStructs.h"
#include "BPStructs.h"
#include "TextureDecoder.h"
#include "TextureMngr.h"
#include "PixelShaderCache.h"
@ -256,7 +256,7 @@ TextureMngr::TCacheEntry* TextureMngr::Load(int texstage, u32 address, int width
if (address == 0)
return NULL;
TexMode0 &tm0 = sumem.tex[texstage > 3].texMode0[texstage & 3];
TexMode0 &tm0 = bpmem.tex[texstage > 3].texMode0[texstage & 3];
u8 *ptr = g_VideoInitialize.pGetMemoryPointer(address);
int bsw = TexDecoder_GetBlockWidthInTexels(tex_format) - 1;
int bsh = TexDecoder_GetBlockHeightInTexels(tex_format) - 1;

2
Source/Plugins/Plugin_VideoOGL/Src/TextureMngr.h
View file

@ -22,7 +22,7 @@
#include "VideoCommon.h"
#include "GLUtil.h"
#include "SUStructs.h"
#include "BPStructs.h"
class TextureMngr
{

30
Source/Plugins/Plugin_VideoOGL/Src/VertexManager.cpp
View file

@ -26,7 +26,7 @@
#include "Profiler.h"
#include "Render.h"
#include "ImageWrite.h"
#include "SUMemory.h"
#include "BPMemory.h"
#include "TextureMngr.h"
#include "PixelShaderCache.h"
#include "PixelShaderManager.h"
@ -155,8 +155,8 @@ void Flush()
#if defined(_DEBUG) || defined(DEBUGFAST)
PRIM_LOG("frame%d:\n texgen=%d, numchan=%d, dualtex=%d, ztex=%d, cole=%d, alpe=%d, ze=%d", g_Config.iSaveTargetId, xfregs.numTexGens,
xfregs.nNumChans, (int)xfregs.bEnableDualTexTransform, sumem.ztex2.op,
sumem.blendmode.colorupdate, sumem.blendmode.alphaupdate, sumem.zmode.updateenable);
xfregs.nNumChans, (int)xfregs.bEnableDualTexTransform, bpmem.ztex2.op,
bpmem.blendmode.colorupdate, bpmem.blendmode.alphaupdate, bpmem.zmode.updateenable);
for (int i = 0; i < xfregs.nNumChans; ++i)
{
@ -177,8 +177,8 @@ void Flush()
xfregs.texcoords[i].postmtxinfo.index, xfregs.texcoords[i].postmtxinfo.normalize);
}
PRIM_LOG("pixel: tev=%d, ind=%d, texgen=%d, dstalpha=%d, alphafunc=0x%x", sumem.genMode.numtevstages+1, sumem.genMode.numindstages,
sumem.genMode.numtexgens, (u32)sumem.dstalpha.enable, (sumem.alphaFunc.hex>>16)&0xff);
PRIM_LOG("pixel: tev=%d, ind=%d, texgen=%d, dstalpha=%d, alphafunc=0x%x", bpmem.genMode.numtevstages+1, bpmem.genMode.numindstages,
bpmem.genMode.numtexgens, (u32)bpmem.dstalpha.enable, (bpmem.alphaFunc.hex>>16)&0xff);
#endif
DVSTARTPROFILE();
@ -198,14 +198,14 @@ void Flush()
DVSTARTSUBPROFILE("VertexManager::Flush:textures");
u32 usedtextures = 0;
for (u32 i = 0; i < (u32)sumem.genMode.numtevstages + 1; ++i)
if (sumem.tevorders[i / 2].getEnable(i & 1))
usedtextures |= 1 << sumem.tevorders[i/2].getTexMap(i & 1);
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; ++i)
if (bpmem.tevorders[i / 2].getEnable(i & 1))
usedtextures |= 1 << bpmem.tevorders[i/2].getTexMap(i & 1);
if (sumem.genMode.numindstages > 0)
for (u32 i = 0; i < (u32)sumem.genMode.numtevstages + 1; ++i)
if (sumem.tevind[i].IsActive() && sumem.tevind[i].bt < sumem.genMode.numindstages)
usedtextures |= 1 << sumem.tevindref.getTexMap(sumem.tevind[i].bt);
if (bpmem.genMode.numindstages > 0)
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; ++i)
if (bpmem.tevind[i].IsActive() && bpmem.tevind[i].bt < bpmem.genMode.numindstages)
usedtextures |= 1 << bpmem.tevindref.getTexMap(bpmem.tevind[i].bt);
u32 nonpow2tex = 0;
for (int i = 0; i < 8; i++)
@ -214,7 +214,7 @@ void Flush()
{
glActiveTexture(GL_TEXTURE0 + i);
FourTexUnits &tex = sumem.tex[i >> 2];
FourTexUnits &tex = bpmem.tex[i >> 2];
TextureMngr::TCacheEntry* tentry = TextureMngr::Load(i, (tex.texImage3[i&3].image_base/* & 0x1FFFFF*/) << 5,
tex.texImage0[i&3].width + 1, tex.texImage0[i&3].height + 1,
tex.texImage0[i&3].format, tex.texTlut[i&3].tmem_offset<<9, tex.texTlut[i&3].tlut_format);
@ -284,7 +284,7 @@ void Flush()
}
// run through vertex groups again to set alpha
if (!g_Config.bDstAlphaPass && sumem.dstalpha.enable && sumem.blendmode.alphaupdate)
if (!g_Config.bDstAlphaPass && bpmem.dstalpha.enable && bpmem.blendmode.alphaupdate)
{
ps = PixelShaderCache::GetShader(true);
@ -309,7 +309,7 @@ void Flush()
// restore color mask
Renderer::SetColorMask();
if (sumem.blendmode.blendenable || sumem.blendmode.subtract)
if (bpmem.blendmode.blendenable || bpmem.blendmode.subtract)
glEnable(GL_BLEND);
}

2
Source/Plugins/Plugin_VideoOGL/Src/VertexShaderCache.h
View file

@ -21,7 +21,7 @@
#include <map>
#include <string>
#include "SUMemory.h"
#include "BPMemory.h"
#include "VertexShaderGen.h"
struct VERTEXSHADER

6
Source/Plugins/Plugin_VideoOGL/Src/main.cpp
View file

@ -71,7 +71,7 @@ Make AA apply instantly during gameplay if possible
#include "Fifo.h"
#include "OpcodeDecoding.h"
#include "TextureMngr.h"
#include "SUStructs.h"
#include "BPStructs.h"
#include "VertexLoader.h"
#include "VertexLoaderManager.h"
#include "VertexManager.h"
@ -330,7 +330,7 @@ void DoState(unsigned char **ptr, int mode) {
// Refresh state.
if (mode == PointerWrap::MODE_READ)
{
SUReload();
BPReload();
RecomputeCachedArraybases();
}
}
@ -347,7 +347,7 @@ void Video_Prepare(void)
TextureMngr::Init();
SUInit();
BPInit();
VertexManager::Init();
Fifo_Init(); // must be done before OpcodeDecoder_Init()
OpcodeDecoder_Init();

2
Source/Plugins/Plugin_VideoOGL/Src/rasterfont.cpp
View file

@ -141,7 +141,7 @@ RasterFont::~RasterFont()
void RasterFont::printString(const char *s, double x, double y, double z)
{
int length = (int)strlen(s);
int length = strlen(s);
if (!length)
return;