ShuriZma
/
dolphin
mirror of https://github.com/dolphin-emu/dolphin.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

THIS BREAKS THE D3D PLUGIN FOR THE NEAR TERM. Resurrect an old patch that moves D3D over to the common shader generator framework. Needs a lot more work. 

git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@2484 8ced0084-cf51-0410-be5f-012b33b47a6e
This commit is contained in:
hrydgard 2009-02-28 22:10:38 +00:00
parent ee44b2a639
commit ecbfec2a13
49 changed files with 1087 additions and 4262 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
Source/Core/VideoCommon/Src/BPMemory.cpp
View File

@ -22,3 +22,42 @@
//BP state
// STATE_TO_SAVE
BPMemory bpmem;
// The plugin must implement this.
void BPWritten(int addr, int changes, int newval);
// 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;
//reset the mask register
if (opcode != 0xFE)
bpmem.bpMask = 0xFFFFFF;
BPWritten(opcode, changes, newval);
}
// 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:
BPWritten(i, 0xFFFFFF, ((u32*)&bpmem)[i]);
}
}
}

3
Source/Core/VideoCommon/Src/NativeVertexFormat.h
View File

@ -96,10 +96,13 @@ public:
virtual void SetupVertexPointers() const = 0;
virtual void EnableComponents(u32 components) {}
int GetVertexStride() const { return vertex_stride; }
static NativeVertexFormat *Create();
// TODO: move these in under private:
u32 m_components; // VB_HAS_X. Bitmask telling what vertex components are present.
u32 vertex_stride;
protected:
// Let subclasses construct.

2
Source/Core/VideoCommon/Src/OpcodeDecoding.cpp
View File

@ -261,7 +261,7 @@ static void Decode()
break;
case GX_CMD_INVL_VC:// Invalidate (vertex cache?)
DebugLog("Invalidate (vertex cache?)");
DEBUG_LOG("Invalidate (vertex cache?)");
break;
case GX_LOAD_BP_REG: //0x61

4
Source/Core/VideoCommon/Src/PixelShaderGen.cpp
View File

@ -928,11 +928,11 @@ static bool WriteAlphaTest(char *&p)
break;
}
WRITE(p, "discard( ");
WRITE(p, "clip( ");
WriteAlphaCompare(p, 0, bpmem.alphaFunc.comp0);
// negated because testing the inverse condition
switch(bpmem.alphaFunc.logic) {
switch (bpmem.alphaFunc.logic) {
case 0: WRITE(p, " || "); break; // and
case 1: WRITE(p, " && "); break; // or
case 2: WRITE(p, " == "); break; // xor

3
Source/Core/VideoCommon/Src/SConscript
View File

@ -5,12 +5,15 @@ Import('env')
files = [
'BPMemory.cpp',
'CPMemory.cpp',
'XFMemory.cpp',
'XFStructs.cpp',
'memcpy_amd.cpp',
'LookUpTables.cpp',
'OpcodeDecoding.cpp',
'TextureDecoder.cpp',
'XFMemory.cpp',
'XFBConvert.cpp',
'IndexGenerator.cpp',
'PixelShaderGen.cpp',
'PixelShaderManager.cpp',
'VertexShaderGen.cpp',

62
Source/Core/VideoCommon/Src/VideoLog.cpp Normal file
View File

@ -0,0 +1,62 @@
// 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 "VideoCommon.h"
#include <stdio.h>
#include <malloc.h>
static FILE* pfLog = NULL;
void __Log(const char *fmt, ...)
{
char* Msg = (char*)alloca(strlen(fmt)+512);
va_list ap;
va_start( ap, fmt );
vsnprintf( Msg, strlen(fmt)+512, fmt, ap );
va_end( ap );
g_VideoInitialize.pLog(Msg, FALSE);
if (pfLog == NULL)
pfLog = fopen(FULL_LOGS_DIR "oglgfx.txt", "w");
if (pfLog != NULL)
fwrite(Msg, strlen(Msg), 1, pfLog);
#ifdef _WIN32
// DWORD tmp;
// WriteConsole(hConsole, Msg, (DWORD)strlen(Msg), &tmp, 0);
#else
//printf("%s", Msg);
#endif
}
void __Log(int type, const char *fmt, ...)
{
char* Msg = (char*)alloca(strlen(fmt)+512);
va_list ap;
va_start( ap, fmt );
vsnprintf( Msg, strlen(fmt)+512, fmt, ap );
va_end( ap );
g_VideoInitialize.pLog(Msg, FALSE);
#ifdef _WIN32
// DWORD tmp;
// WriteConsole(hConsole, Msg, (DWORD)strlen(Msg), &tmp, 0);
#endif
}

469
Source/Plugins/Plugin_VideoOGL/Src/XFStructs.cpp → Source/Core/VideoCommon/Src/XFStructs.cpp
View File

@ -1,236 +1,233 @@
// 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 "Globals.h"
#include "XFMemory.h"
#include "VertexManager.h"
#include "VertexShaderManager.h"
#include "PixelShaderManager.h"
// LoadXFReg 0x10
void LoadXFReg(u32 transferSize, u32 baseAddress, u32 *pData)
{
u32 address = baseAddress;
for (int i = 0; i < (int)transferSize; i++)
{
address = baseAddress + i;
// Setup a Matrix
if (address < 0x1000)
{
VertexManager::Flush();
VertexShaderManager::InvalidateXFRange(address, address + transferSize);
//PRIM_LOG("xfmem write: 0x%x-0x%x\n", address, address+transferSize);
u32* p1 = &xfmem[address];
memcpy_gc(p1, &pData[i], transferSize*4);
i += transferSize;
}
else if (address<0x2000)
{
u32 data = pData[i];
switch (address)
{
case 0x1000: // error
break;
case 0x1001: // diagnostics
break;
case 0x1002: // internal state 0
break;
case 0x1003: // internal state 1
break;
case 0x1004: // xf_clock
break;
case 0x1005: // clipdisable
if (data & 1) { // disable clipping detection
}
if (data & 2) { // disable trivial rejection
}
if (data & 4) { // disable cpoly clipping acceleration
}
break;
case 0x1006: //SetGPMetric
break;
case 0x1008: //__GXXfVtxSpecs, wrote 0004
xfregs.hostinfo = *(INVTXSPEC*)&data;
break;
case 0x1009: //GXSetNumChans (no)
if ((u32)xfregs.nNumChans != (data&3)) {
VertexManager::Flush();
xfregs.nNumChans = data&3;
}
break;
case 0x100a: //GXSetChanAmbientcolor
if (xfregs.colChans[0].ambColor != data) {
VertexManager::Flush();
xfregs.colChans[0].ambColor = data;
VertexShaderManager::SetMaterialColor(0, data);
}
break;
case 0x100b: //GXSetChanAmbientcolor
if (xfregs.colChans[1].ambColor != data) {
VertexManager::Flush();
xfregs.colChans[1].ambColor = data;
VertexShaderManager::SetMaterialColor(1, data);
}
break;
case 0x100c: //GXSetChanMatcolor (rgba)
if (xfregs.colChans[0].matColor != data) {
VertexManager::Flush();
xfregs.colChans[0].matColor = data;
VertexShaderManager::SetMaterialColor(2, data);
}
break;
case 0x100d: //GXSetChanMatcolor (rgba)
if (xfregs.colChans[1].matColor != data) {
VertexManager::Flush();
xfregs.colChans[1].matColor = data;
VertexShaderManager::SetMaterialColor(3, data);
}
break;
case 0x100e: // color0
if (xfregs.colChans[0].color.hex != (data & 0x7fff) ) {
VertexManager::Flush();
xfregs.colChans[0].color.hex = data;
}
break;
case 0x100f: // color1
if (xfregs.colChans[1].color.hex != (data & 0x7fff) ) {
VertexManager::Flush();
xfregs.colChans[1].color.hex = data;
}
break;
case 0x1010: // alpha0
if (xfregs.colChans[0].alpha.hex != (data & 0x7fff) ) {
VertexManager::Flush();
xfregs.colChans[0].alpha.hex = data;
}
break;
case 0x1011: // alpha1
if (xfregs.colChans[1].alpha.hex != (data & 0x7fff) ) {
VertexManager::Flush();
xfregs.colChans[1].alpha.hex = data;
}
break;
case 0x1012: // dual tex transform
if (xfregs.bEnableDualTexTransform != (data & 1)) {
VertexManager::Flush();
xfregs.bEnableDualTexTransform = data & 1;
}
break;
case 0x1013:
case 0x1014:
case 0x1015:
case 0x1016:
case 0x1017:
DEBUG_LOG(VIDEO, "xf addr: %x=%x\n", address, data);
break;
case 0x1018:
//_assert_msg_(GX_XF, 0, "XF matrixindex0");
VertexShaderManager::SetTexMatrixChangedA(data); //?
break;
case 0x1019:
//_assert_msg_(GX_XF, 0, "XF matrixindex1");
VertexShaderManager::SetTexMatrixChangedB(data); //?
break;
case 0x101a:
VertexManager::Flush();
VertexShaderManager::SetViewport((float*)&pData[i]);
PixelShaderManager::SetViewport((float*)&pData[i]);
i += 6;
break;
case 0x101c: // paper mario writes 16777216.0f, 1677721.75
break;
case 0x101f: // paper mario writes 16777216.0f, 5033165.0f
break;
case 0x1020:
VertexManager::Flush();
VertexShaderManager::SetProjection((float*)&pData[i]);
i += 7;
return;
case 0x103f: // GXSetNumTexGens
if ((u32)xfregs.numTexGens != data) {
VertexManager::Flush();
xfregs.numTexGens = data;
}
break;
case 0x1040: xfregs.texcoords[0].texmtxinfo.hex = data; break;
case 0x1041: xfregs.texcoords[1].texmtxinfo.hex = data; break;
case 0x1042: xfregs.texcoords[2].texmtxinfo.hex = data; break;
case 0x1043: xfregs.texcoords[3].texmtxinfo.hex = data; break;
case 0x1044: xfregs.texcoords[4].texmtxinfo.hex = data; break;
case 0x1045: xfregs.texcoords[5].texmtxinfo.hex = data; break;
case 0x1046: xfregs.texcoords[6].texmtxinfo.hex = data; break;
case 0x1047: xfregs.texcoords[7].texmtxinfo.hex = data; break;
case 0x1048:
case 0x1049:
case 0x104a:
case 0x104b:
case 0x104c:
case 0x104d:
case 0x104e:
case 0x104f:
DEBUG_LOG(VIDEO, "xf addr: %x=%x\n", address, data);
break;
case 0x1050: xfregs.texcoords[0].postmtxinfo.hex = data; break;
case 0x1051: xfregs.texcoords[1].postmtxinfo.hex = data; break;
case 0x1052: xfregs.texcoords[2].postmtxinfo.hex = data; break;
case 0x1053: xfregs.texcoords[3].postmtxinfo.hex = data; break;
case 0x1054: xfregs.texcoords[4].postmtxinfo.hex = data; break;
case 0x1055: xfregs.texcoords[5].postmtxinfo.hex = data; break;
case 0x1056: xfregs.texcoords[6].postmtxinfo.hex = data; break;
case 0x1057: xfregs.texcoords[7].postmtxinfo.hex = data; break;
default:
DEBUG_LOG(VIDEO, "xf addr: %x=%x\n", address, data);
break;
}
}
else if (address >= 0x4000)
{
// MessageBox(NULL, "1", "1", MB_OK);
//4010 __GXSetGenMode
}
}
}
// TODO - verify that it is correct. Seems to work, though.
void LoadIndexedXF(u32 val, int array)
{
int index = val >> 16;
int address = val & 0xFFF; //check mask
int size = ((val >> 12) & 0xF) + 1;
//load stuff from array to address in xf mem
VertexManager::Flush();
VertexShaderManager::InvalidateXFRange(address, address + size);
//PRIM_LOG("xfmem iwrite: 0x%x-0x%x\n", address, address + size);
for (int i = 0; i < size; i++)
xfmem[address + i] = Memory_Read_U32(arraybases[array] + arraystrides[array] * index + i * 4);
}
// 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 "VideoCommon.h"
#include "XFMemory.h"
#include "CPMemory.h"
#include "NativeVertexWriter.h"
#include "VertexShaderManager.h"
// LoadXFReg 0x10
void LoadXFReg(u32 transferSize, u32 baseAddress, u32 *pData)
{
u32 address = baseAddress;
for (int i = 0; i < (int)transferSize; i++)
{
address = baseAddress + i;
// Setup a Matrix
if (address < 0x1000)
{
VertexManager::Flush();
VertexShaderManager::InvalidateXFRange(address, address + transferSize);
//PRIM_LOG("xfmem write: 0x%x-0x%x\n", address, address+transferSize);
u32* p1 = &xfmem[address];
memcpy_gc(p1, &pData[i], transferSize*4);
i += transferSize;
}
else if (address < 0x2000)
{
u32 data = pData[i];
switch (address)
{
case 0x1000: // error
break;
case 0x1001: // diagnostics
break;
case 0x1002: // internal state 0
break;
case 0x1003: // internal state 1
break;
case 0x1004: // xf_clock
break;
case 0x1005: // clipdisable
if (data & 1) { // disable clipping detection
}
if (data & 2) { // disable trivial rejection
}
if (data & 4) { // disable cpoly clipping acceleration
}
break;
case 0x1006: //SetGPMetric
break;
case 0x1008: //__GXXfVtxSpecs, wrote 0004
xfregs.hostinfo = *(INVTXSPEC*)&data;
break;
case 0x1009: //GXSetNumChans (no)
if ((u32)xfregs.nNumChans != (data & 3)) {
VertexManager::Flush();
xfregs.nNumChans = data & 3;
}
break;
case 0x100a: //GXSetChanAmbientcolor
if (xfregs.colChans[0].ambColor != data) {
VertexManager::Flush();
xfregs.colChans[0].ambColor = data;
VertexShaderManager::SetMaterialColor(0, data);
}
break;
case 0x100b: //GXSetChanAmbientcolor
if (xfregs.colChans[1].ambColor != data) {
VertexManager::Flush();
xfregs.colChans[1].ambColor = data;
VertexShaderManager::SetMaterialColor(1, data);
}
break;
case 0x100c: //GXSetChanMatcolor (rgba)
if (xfregs.colChans[0].matColor != data) {
VertexManager::Flush();
xfregs.colChans[0].matColor = data;
VertexShaderManager::SetMaterialColor(2, data);
}
break;
case 0x100d: //GXSetChanMatcolor (rgba)
if (xfregs.colChans[1].matColor != data) {
VertexManager::Flush();
xfregs.colChans[1].matColor = data;
VertexShaderManager::SetMaterialColor(3, data);
}
break;
case 0x100e: // color0
if (xfregs.colChans[0].color.hex != (data & 0x7fff) ) {
VertexManager::Flush();
xfregs.colChans[0].color.hex = data;
}
break;
case 0x100f: // color1
if (xfregs.colChans[1].color.hex != (data & 0x7fff) ) {
VertexManager::Flush();
xfregs.colChans[1].color.hex = data;
}
break;
case 0x1010: // alpha0
if (xfregs.colChans[0].alpha.hex != (data & 0x7fff) ) {
VertexManager::Flush();
xfregs.colChans[0].alpha.hex = data;
}
break;
case 0x1011: // alpha1
if (xfregs.colChans[1].alpha.hex != (data & 0x7fff) ) {
VertexManager::Flush();
xfregs.colChans[1].alpha.hex = data;
}
break;
case 0x1012: // dual tex transform
if (xfregs.bEnableDualTexTransform != (data & 1)) {
VertexManager::Flush();
xfregs.bEnableDualTexTransform = data & 1;
}
break;
case 0x1013:
case 0x1014:
case 0x1015:
case 0x1016:
case 0x1017:
DEBUG_LOG("xf addr: %x=%x\n", address, data);
break;
case 0x1018:
//_assert_msg_(GX_XF, 0, "XF matrixindex0");
VertexShaderManager::SetTexMatrixChangedA(data); //?
break;
case 0x1019:
//_assert_msg_(GX_XF, 0, "XF matrixindex1");
VertexShaderManager::SetTexMatrixChangedB(data); //?
break;
case 0x101a:
VertexManager::Flush();
VertexShaderManager::SetViewport((float*)&pData[i]);
i += 6;
break;
case 0x101c: // paper mario writes 16777216.0f, 1677721.75
break;
case 0x101f: // paper mario writes 16777216.0f, 5033165.0f
break;
case 0x1020:
VertexManager::Flush();
VertexShaderManager::SetProjection((float*)&pData[i]);
i += 7;
return;
case 0x103f: // GXSetNumTexGens
if ((u32)xfregs.numTexGens != data) {
VertexManager::Flush();
xfregs.numTexGens = data;
}
break;
case 0x1040: xfregs.texcoords[0].texmtxinfo.hex = data; break;
case 0x1041: xfregs.texcoords[1].texmtxinfo.hex = data; break;
case 0x1042: xfregs.texcoords[2].texmtxinfo.hex = data; break;
case 0x1043: xfregs.texcoords[3].texmtxinfo.hex = data; break;
case 0x1044: xfregs.texcoords[4].texmtxinfo.hex = data; break;
case 0x1045: xfregs.texcoords[5].texmtxinfo.hex = data; break;
case 0x1046: xfregs.texcoords[6].texmtxinfo.hex = data; break;
case 0x1047: xfregs.texcoords[7].texmtxinfo.hex = data; break;
case 0x1048:
case 0x1049:
case 0x104a:
case 0x104b:
case 0x104c:
case 0x104d:
case 0x104e:
case 0x104f:
DEBUG_LOG("xf addr: %x=%x\n", address, data);
break;
case 0x1050: xfregs.texcoords[0].postmtxinfo.hex = data; break;
case 0x1051: xfregs.texcoords[1].postmtxinfo.hex = data; break;
case 0x1052: xfregs.texcoords[2].postmtxinfo.hex = data; break;
case 0x1053: xfregs.texcoords[3].postmtxinfo.hex = data; break;
case 0x1054: xfregs.texcoords[4].postmtxinfo.hex = data; break;
case 0x1055: xfregs.texcoords[5].postmtxinfo.hex = data; break;
case 0x1056: xfregs.texcoords[6].postmtxinfo.hex = data; break;
case 0x1057: xfregs.texcoords[7].postmtxinfo.hex = data; break;
default:
DEBUG_LOG("xf addr: %x=%x\n", address, data);
break;
}
}
else if (address >= 0x4000)
{
// MessageBox(NULL, "1", "1", MB_OK);
//4010 __GXSetGenMode
}
}
}
// TODO - verify that it is correct. Seems to work, though.
void LoadIndexedXF(u32 val, int array)
{
int index = val >> 16;
int address = val & 0xFFF; //check mask
int size = ((val >> 12) & 0xF) + 1;
//load stuff from array to address in xf mem
VertexManager::Flush();
VertexShaderManager::InvalidateXFRange(address, address+size);
//PRIM_LOG("xfmem iwrite: 0x%x-0x%x\n", address, address+size);
for (int i = 0; i < size; i++)
xfmem[address + i] = Memory_Read_U32(arraybases[array] + arraystrides[array]*index + i*4);
}

36
Source/Plugins/Plugin_VideoOGL/Src/XFStructs.h → Source/Core/VideoCommon/Src/XFStructs.h
View File

@ -1,18 +1,18 @@
// 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 "XFMemory.h"
// 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 "XFMemory.h"

28
Source/Core/VideoCommon/VideoCommon.vcproj
View File

@ -458,6 +458,14 @@
RelativePath=".\Src\XFMemory.h"
>
</File>
<File
RelativePath=".\Src\XFStructs.cpp"
>
</File>
<File
RelativePath=".\Src\XFStructs.h"
>
</File>
</Filter>
<Filter
Name="ShaderManagers"
@ -470,6 +478,14 @@
RelativePath=".\Src\PixelShaderManager.h"
>
</File>
<File
RelativePath=".\Src\VertexShaderManager.cpp"
>
</File>
<File
RelativePath=".\Src\VertexShaderManager.h"
>
</File>
</Filter>
<Filter
Name="Util"
@ -643,18 +659,14 @@
RelativePath=".\Src\VertexLoaderManager.h"
>
</File>
<File
RelativePath=".\Src\VertexShaderManager.cpp"
>
</File>
<File
RelativePath=".\Src\VertexShaderManager.h"
>
</File>
<File
RelativePath=".\Src\VideoCommon.h"
>
</File>
<File
RelativePath=".\Src\VideoLog.cpp"
>
</File>
</Files>
<Globals>
</Globals>

6
Source/Plugins/Plugin_DSP_HLE/Src/main.cpp
View File

@ -250,9 +250,9 @@ void Initialize(void *init)
}
#if defined(WIN32) && defined(_DEBUG)
int tmpflag = _CrtSetDbgFlag(_CRTDBG_REPORT_FLAG);
tmpflag |= _CRTDBG_DELAY_FREE_MEM_DF;
_CrtSetDbgFlag(tmpflag);
//int tmpflag = _CrtSetDbgFlag(_CRTDBG_REPORT_FLAG);
//tmpflag |= _CRTDBG_DELAY_FREE_MEM_DF;
//_CrtSetDbgFlag(tmpflag);
#endif
if (soundStream)

88
Source/Plugins/Plugin_VideoDX9/Plugin_VideoDX9.vcproj
View File

@ -1185,98 +1185,18 @@
RelativePath=".\Src\BPStructs.h"
>
</File>
<File
RelativePath=".\Src\CPStructs.cpp"
>
</File>
<File
RelativePath=".\Src\CPStructs.h"
>
</File>
<File
RelativePath=".\Src\DecodedVArray.cpp"
>
</File>
<File
RelativePath=".\Src\DecodedVArray.h"
>
</File>
<File
RelativePath=".\Src\NativeVertexFormat.cpp"
>
</File>
<File
RelativePath=".\Src\TransformEngine.cpp"
>
</File>
<File
RelativePath=".\Src\TransformEngine.h"
>
</File>
<File
RelativePath=".\Src\VertexLoader.cpp"
>
</File>
<File
RelativePath=".\Src\VertexLoader.h"
>
</File>
<File
RelativePath=".\Src\VertexLoader_Color.h"
>
</File>
<File
RelativePath=".\Src\VertexLoader_MtxIndex.h"
>
</File>
<File
RelativePath=".\Src\VertexLoader_Normal.cpp"
>
</File>
<File
RelativePath=".\Src\VertexLoader_Normal.h"
>
</File>
<File
RelativePath=".\Src\VertexLoader_Position.h"
>
</File>
<File
RelativePath=".\Src\VertexLoader_TextCoord.h"
>
</File>
<File
RelativePath=".\Src\VertexLoaderManager.cpp"
>
</File>
<File
RelativePath=".\Src\VertexManager.cpp"
>
</File>
<File
RelativePath=".\Src\VertexManager.h"
>
</File>
<File
RelativePath=".\Src\XFStructs.cpp"
>
</File>
<File
RelativePath=".\Src\XFStructs.h"
RelativePath=".\Src\VertexShaderCache.cpp"
>
</File>
</Filter>
<Filter
Name="Render"
>
<File
RelativePath=".\Src\PixelShader.cpp"
>
</File>
<File
RelativePath=".\Src\PixelShader.h"
>
</File>
<File
RelativePath=".\Src\PixelShaderCache.cpp"
>
@ -1302,15 +1222,15 @@
>
</File>
<File
RelativePath=".\Src\VertexShader.cpp"
RelativePath=".\Src\VertexLoaderManager.cpp"
>
</File>
<File
RelativePath=".\Src\VertexShader.h"
RelativePath=".\Src\VertexManager.cpp"
>
</File>
<File
RelativePath=".\Src\VertexShaderCache.cpp"
RelativePath=".\Src\VertexManager.h"
>
</File>
<File

649
Source/Plugins/Plugin_VideoDX9/Src/BPStructs.cpp
View File

@ -25,7 +25,8 @@
#include "TextureCache.h"
#include "TextureDecoder.h"
#include "VertexManager.h"
#include "PixelShader.h"
#include "PixelShaderGen.h"
#include "PixelShaderManager.h"
#include "Utils.h"
#include "main.h" //for the plugin interface
@ -34,8 +35,10 @@ bool textureChanged[8];
const bool renderFog = false;
using namespace D3D;
// State translation lookup tables
const D3DBLEND d3dSrcFactors[8] =
static const D3DBLEND d3dSrcFactors[8] =
{
D3DBLEND_ZERO,
D3DBLEND_ONE,
@ -47,7 +50,7 @@ const D3DBLEND d3dSrcFactors[8] =
D3DBLEND_INVDESTALPHA
};
const D3DBLEND d3dDestFactors[8] =
static const D3DBLEND d3dDestFactors[8] =
{
D3DBLEND_ZERO,
D3DBLEND_ONE,
@ -59,7 +62,7 @@ const D3DBLEND d3dDestFactors[8] =
D3DBLEND_INVDESTALPHA
};
const D3DCULL d3dCullModes[4] =
static const D3DCULL d3dCullModes[4] =
{
D3DCULL_NONE,
D3DCULL_CCW,
@ -67,7 +70,7 @@ const D3DCULL d3dCullModes[4] =
D3DCULL_CCW
};
const D3DCMPFUNC d3dCmpFuncs[8] =
static const D3DCMPFUNC d3dCmpFuncs[8] =
{
D3DCMP_NEVER,
D3DCMP_LESS,
@ -79,7 +82,7 @@ const D3DCMPFUNC d3dCmpFuncs[8] =
D3DCMP_ALWAYS
};
const D3DTEXTUREFILTERTYPE d3dMipFilters[4] =
static const D3DTEXTUREFILTERTYPE d3dMipFilters[4] =
{
D3DTEXF_NONE,
D3DTEXF_POINT,
@ -87,7 +90,7 @@ const D3DTEXTUREFILTERTYPE d3dMipFilters[4] =
D3DTEXF_LINEAR, //reserved
};
const D3DTEXTUREADDRESS d3dClamps[4] =
static const D3DTEXTUREADDRESS d3dClamps[4] =
{
D3DTADDRESS_CLAMP,
D3DTADDRESS_WRAP,
@ -101,7 +104,6 @@ void BPInit()
bpmem.bpMask = 0xFFFFFF;
}
using namespace D3D;
// __________________________________________________________________________________________________
// BPWritten
//
@ -116,12 +118,12 @@ void BPWritten(int addr, int changes, int newval)
((u32*)&bpmem)[addr] = newval;
// dev->SetRenderState(D3DRS_CULLMODE, d3dCullModes[bpmem.genMode.cullmode]);
Renderer::SetRenderState( D3DRS_CULLMODE, d3dCullModes[bpmem.genMode.cullmode] );
Renderer::SetRenderState(D3DRS_CULLMODE, d3dCullModes[bpmem.genMode.cullmode]);
if (bpmem.genMode.cullmode == 3)
{
// dev->SetRenderState(D3DRS_COLORWRITEENABLE, 0);
Renderer::SetRenderState( D3DRS_COLORWRITEENABLE, 0 );
Renderer::SetRenderState(D3DRS_COLORWRITEENABLE, 0);
}
else
{
@ -132,7 +134,7 @@ void BPWritten(int addr, int changes, int newval)
write |= D3DCOLORWRITEENABLE_RED | D3DCOLORWRITEENABLE_GREEN | D3DCOLORWRITEENABLE_BLUE;
// dev->SetRenderState(D3DRS_COLORWRITEENABLE, write);
Renderer::SetRenderState( D3DRS_COLORWRITEENABLE, write );
Renderer::SetRenderState(D3DRS_COLORWRITEENABLE, write);
}
}
break;
@ -148,7 +150,7 @@ void BPWritten(int addr, int changes, int newval)
if (changes) {
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
// PixelShaderMngr::SetIndMatrixChanged((addr-BPMEM_IND_MTX)/3);
PixelShaderManager::SetIndMatrixChanged((addr-BPMEM_IND_MTX)/3);
}
break;
case BPMEM_RAS1_SS0:
@ -156,7 +158,7 @@ void BPWritten(int addr, int changes, int newval)
if (changes) {
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
// PixelShaderMngr::SetIndTexScaleChanged();
PixelShaderManager::SetIndTexScaleChanged();
}
break;
@ -171,9 +173,9 @@ void BPWritten(int addr, int changes, int newval)
// dev->SetRenderState(D3DRS_ZWRITEENABLE, bpmem.zmode.updateenable);
// dev->SetRenderState(D3DRS_ZFUNC,d3dCmpFuncs[bpmem.zmode.func]);
Renderer::SetRenderState( D3DRS_ZENABLE, TRUE );
Renderer::SetRenderState( D3DRS_ZWRITEENABLE, bpmem.zmode.updateenable );
Renderer::SetRenderState( D3DRS_ZFUNC, d3dCmpFuncs[bpmem.zmode.func] );
Renderer::SetRenderState(D3DRS_ZENABLE, TRUE);
Renderer::SetRenderState(D3DRS_ZWRITEENABLE, bpmem.zmode.updateenable);
Renderer::SetRenderState(D3DRS_ZFUNC, d3dCmpFuncs[bpmem.zmode.func]);
}
else
{
@ -181,50 +183,33 @@ void BPWritten(int addr, int changes, int newval)
// dev->SetRenderState(D3DRS_ZENABLE, FALSE);
// dev->SetRenderState(D3DRS_ZWRITEENABLE, FALSE);
Renderer::SetRenderState( D3DRS_ZENABLE, FALSE );
Renderer::SetRenderState( D3DRS_ZWRITEENABLE, FALSE );
Renderer::SetRenderState(D3DRS_ZENABLE, FALSE);
Renderer::SetRenderState(D3DRS_ZWRITEENABLE, FALSE);
}
//if (!bpmem.zmode.updateenable)
// Renderer::SetRenderMode(Renderer::RM_Normal);
}
break;
case BPMEM_ALPHACOMPARE:
if (changes)
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
float f[4] =
{
bpmem.alphaFunc.ref0/255.0f,
bpmem.alphaFunc.ref1/255.0f,
0,0
};
dev->SetPixelShaderConstantF(PS_CONST_ALPHAREF,f,1);
if (D3D::GetShaderVersion() == PSNONE)
{
// dev->SetRenderState(D3DRS_ALPHATESTENABLE, (Compare)bpmem.alphaFunc.comp0 != COMPARE_ALWAYS);
// dev->SetRenderState(D3DRS_ALPHAREF, bpmem.alphaFunc.ref0*4);
// dev->SetRenderState(D3DRS_ALPHAFUNC, d3dCmpFuncs[bpmem.alphaFunc.comp0]);
Renderer::SetRenderState( D3DRS_ALPHATESTENABLE, (Compare)bpmem.alphaFunc.comp0 != COMPARE_ALWAYS );
Renderer::SetRenderState( D3DRS_ALPHAREF, bpmem.alphaFunc.ref0 * 4 );
Renderer::SetRenderState( D3DRS_ALPHAFUNC, d3dCmpFuncs[bpmem.alphaFunc.comp0] );
}
// Normally, use texkill in pixel shader to emulate alpha test
}
break;
case BPMEM_CONSTANTALPHA:
if (changes)
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
float f[4] = {
bpmem.dstalpha.alpha/255.0f,0,0,0
};
dev->SetPixelShaderConstantF(PS_CONST_CONSTALPHA,f,1);
}
break;
case BPMEM_ALPHACOMPARE:
if (changes) {
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
PRIM_LOG("alphacmp: ref0=%d, ref1=%d, comp0=%d, comp1=%d, logic=%d\n", bpmem.alphaFunc.ref0,
bpmem.alphaFunc.ref1, bpmem.alphaFunc.comp0, bpmem.alphaFunc.comp1, bpmem.alphaFunc.logic);
PixelShaderManager::SetAlpha(bpmem.alphaFunc);
}
break;
case BPMEM_CONSTANTALPHA:
if (changes) {
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
PRIM_LOG("constalpha: alp=%d, en=%d\n", bpmem.dstalpha.alpha, bpmem.dstalpha.enable);
PixelShaderManager::SetDestAlpha(bpmem.dstalpha);
}
break;
case BPMEM_LINEPTWIDTH:
{
@ -251,7 +236,7 @@ void BPWritten(int addr, int changes, int newval)
if (changes & 1)
{
// dev->SetRenderState(D3DRS_ALPHABLENDENABLE,bpmem.blendmode.blendenable);
Renderer::SetRenderState( D3DRS_ALPHABLENDENABLE, bpmem.blendmode.blendenable );
Renderer::SetRenderState(D3DRS_ALPHABLENDENABLE, bpmem.blendmode.blendenable);
}
if (changes & 2) {} // Logic op blending. D3D can't do this but can fake some modes.
if (changes & 4) {
@ -264,18 +249,18 @@ void BPWritten(int addr, int changes, int newval)
if (changes & 0x700)
{
// dev->SetRenderState(D3DRS_SRCBLEND, src);
Renderer::SetRenderState( D3DRS_SRCBLEND, src );
Renderer::SetRenderState(D3DRS_SRCBLEND, src);
}
if (changes & 0xE0) {
if (!bpmem.blendmode.subtract)
{
// dev->SetRenderState(D3DRS_DESTBLEND, dst);
Renderer::SetRenderState( D3DRS_DESTBLEND, dst );
Renderer::SetRenderState(D3DRS_DESTBLEND, dst);
}
else
{
// dev->SetRenderState(D3DRS_DESTBLEND, D3DBLEND_ONE);
Renderer::SetRenderState( D3DRS_DESTBLEND, D3DBLEND_ONE );
Renderer::SetRenderState(D3DRS_DESTBLEND, D3DBLEND_ONE);
}
}
if (changes & 0x800) {
@ -283,20 +268,20 @@ void BPWritten(int addr, int changes, int newval)
{
// dev->SetRenderState(D3DRS_SRCBLEND, D3DBLEND_ONE);
// dev->SetRenderState(D3DRS_DESTBLEND, D3DBLEND_ONE);
Renderer::SetRenderState( D3DRS_SRCBLEND, D3DBLEND_ONE );
Renderer::SetRenderState( D3DRS_DESTBLEND, D3DBLEND_ONE );
Renderer::SetRenderState(D3DRS_SRCBLEND, D3DBLEND_ONE);
Renderer::SetRenderState(D3DRS_DESTBLEND, D3DBLEND_ONE);
}
else
{
// dev->SetRenderState(D3DRS_SRCBLEND, src);
// dev->SetRenderState(D3DRS_DESTBLEND, dst);
Renderer::SetRenderState( D3DRS_SRCBLEND, src );
Renderer::SetRenderState( D3DRS_DESTBLEND, dst );
Renderer::SetRenderState(D3DRS_SRCBLEND, src);
Renderer::SetRenderState(D3DRS_DESTBLEND, dst);
}
// dev->SetRenderState(D3DRS_BLENDOP,bpmem.blendmode.subtract?D3DBLENDOP_SUBTRACT:D3DBLENDOP_ADD);
Renderer::SetRenderState( D3DRS_BLENDOP, bpmem.blendmode.subtract ? D3DBLENDOP_SUBTRACT : D3DBLENDOP_ADD );
Renderer::SetRenderState(D3DRS_BLENDOP, bpmem.blendmode.subtract ? D3DBLENDOP_SUBTRACT : D3DBLENDOP_ADD);
}
//if (bpmem.blendmode.logicopenable) // && bpmem.blendmode.logicmode == 4)
// MessageBox(0,"LOGIC",0,0);
@ -311,13 +296,13 @@ void BPWritten(int addr, int changes, int newval)
write |= D3DCOLORWRITEENABLE_RED | D3DCOLORWRITEENABLE_GREEN | D3DCOLORWRITEENABLE_BLUE;
// dev->SetRenderState(D3DRS_COLORWRITEENABLE, write);
Renderer::SetRenderState( D3DRS_COLORWRITEENABLE, write );
Renderer::SetRenderState(D3DRS_COLORWRITEENABLE, write);
}
}
break;
case BPMEM_FOGRANGE:
if(changes) {
if (changes) {
// TODO(XK): Fog range format
//Renderer::SetRenderState(D3DRS_FOGSTART, ...
//Renderer::SetRenderState(D3DRS_FOGEND, ...
@ -343,12 +328,12 @@ void BPWritten(int addr, int changes, int newval)
case BPMEM_FOGPARAM3:
//fog settings
if(changes) {
if (changes) {
static bool bFog = false;
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
if(!renderFog)
if (!renderFog)
break;
/// u32 fogCTemp = bpmem.fog.c_proj_fsel.cShifted12 << 12;
@ -359,7 +344,7 @@ void BPWritten(int addr, int changes, int newval)
switch(bpmem.fog.c_proj_fsel.fsel)
{
case 0: // Off
if(bFog) {
if (bFog) {
Renderer::SetRenderState(D3DRS_FOGENABLE, false);
bFog = false;
}
@ -383,7 +368,7 @@ void BPWritten(int addr, int changes, int newval)
break;
}
if(bpmem.fog.c_proj_fsel.fsel > 0 && !bFog) {
if (bpmem.fog.c_proj_fsel.fsel > 0 && !bFog) {
Renderer::SetRenderState(D3DRS_FOGENABLE, true);
bFog = true;
}
@ -392,11 +377,11 @@ void BPWritten(int addr, int changes, int newval)
break;
case BPMEM_FOGCOLOR:
if(changes) {
if (changes) {
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
if(!renderFog)
if (!renderFog)
break;
// dev->SetRenderState(D3DRS_FOGCOLOR,bpmem.fog.color);
@ -414,35 +399,10 @@ void BPWritten(int addr, int changes, int newval)
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
Renderer::SetScissorRect();
}
break;
case BPMEM_SCISSORTL:
case BPMEM_SCISSORBR:
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
int xoff = bpmem.scissorOffset.x*2-342;
int yoff = bpmem.scissorOffset.y*2-342;
RECT rc;
rc.left=bpmem.scissorTL.x + xoff - 342 -1;
if (rc.left<0) rc.left=0;
rc.top=bpmem.scissorTL.y + yoff - 342 -1;
if (rc.top<0) rc.top=0;
rc.right=bpmem.scissorBR.x + xoff - 342 +2;
if (rc.right>640) rc.right=640;
rc.bottom=bpmem.scissorBR.y + yoff - 342 +2;
if (rc.bottom>480) rc.bottom=480;
char temp[256];
sprintf(temp,"ScissorRect: %i %i %i %i",rc.left,rc.top,rc.right,rc.bottom);
g_VideoInitialize.pLog(temp, FALSE);
// dev->SetRenderState(D3DRS_SCISSORTESTENABLE,TRUE);
Renderer::SetRenderState( D3DRS_SCISSORTESTENABLE, TRUE );
Renderer::SetScissorBox(rc);
}
break;
case BPMEM_ZTEX1:
if (changes) {
VertexManager::Flush();
@ -458,11 +418,146 @@ void BPWritten(int addr, int changes, int newval)
#ifdef _DEBUG
const char* pzop[] = {"DISABLE", "ADD", "REPLACE", "?"};
const char* pztype[] = {"Z8", "Z16", "Z24", "?"};
DebugLog("ztex op=%s, type=%s\n", pzop[bpmem.ztex2.op], pztype[bpmem.ztex2.type]);
DEBUG_LOG("ztex op=%s, type=%s\n", pzop[bpmem.ztex2.op], pztype[bpmem.ztex2.type]);
#endif
}
break;
case 0x45: //GXSetDrawDone
VertexManager::Flush();
switch (newval & 0xFF)
{
case 0x02:
g_VideoInitialize.pSetPEFinish(); // may generate interrupt
DEBUG_LOG("GXSetDrawDone SetPEFinish (value: 0x%02X)", (newval & 0xFFFF));
break;
default:
DEBUG_LOG("GXSetDrawDone ??? (value 0x%02X)", (newval & 0xFFFF));
break;
}
break;
case BPMEM_PE_TOKEN_ID:
g_VideoInitialize.pSetPEToken(static_cast<u16>(newval & 0xFFFF), FALSE);
DEBUG_LOG("SetPEToken 0x%04x", (newval & 0xFFFF));
break;
case BPMEM_PE_TOKEN_INT_ID:
g_VideoInitialize.pSetPEToken(static_cast<u16>(newval & 0xFFFF), TRUE);
DEBUG_LOG("SetPEToken + INT 0x%04x", (newval & 0xFFFF));
break;
case 0x67: // set gp metric?
break;
case BPMEM_TRIGGER_EFB_COPY:
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
RECT rc = {
(LONG)(bpmem.copyTexSrcXY.x*Renderer::GetXScale()),
(LONG)(bpmem.copyTexSrcXY.y*Renderer::GetYScale()),
(LONG)((bpmem.copyTexSrcXY.x+bpmem.copyTexSrcWH.x)*Renderer::GetXScale()),
(LONG)((bpmem.copyTexSrcXY.y+bpmem.copyTexSrcWH.y)*Renderer::GetYScale())
};
UPE_Copy PE_copy;
PE_copy.Hex = bpmem.triggerEFBCopy;
// clamp0
// clamp1
// target_pixel_format
// gamma
// scale_something
// clear
// frame_to_field
// copy_to_xfb
// ???: start Mem to/from EFB transfer
/* bool bMip = false; // ignored
if (bpmem.triggerEFBCopy & EFBCOPY_GENERATEMIPS)
bMip = true;*/
if (PE_copy.copy_to_xfb == 0) // bpmem.triggerEFBCopy & EFBCOPY_EFBTOTEXTURE)
{
// EFB to texture
// for some reason it sets bpmem.zcontrol.pixel_format to PIXELFMT_Z24 every time a zbuffer format is given as a dest to GXSetTexCopyDst
TextureCache::CopyEFBToRenderTarget(bpmem.copyTexDest<<5, &rc);
}
else
{
// EFB to XFB
// MessageBox(0, "WASDF", 0, 0);
Renderer::SwapBuffers();
PRIM_LOG("Renderer::SwapBuffers()");
g_VideoInitialize.pCopiedToXFB();
}
// clearing
if (PE_copy.clear) // bpmem.triggerEFBCopy & EFBCOPY_CLEAR)
{
// 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
// clear color
u32 clearColor = (bpmem.clearcolorAR << 16) | bpmem.clearcolorGB;
if (bpmem.blendmode.colorupdate)
{
D3DRECT drc;
drc.x1 = rc.left;
drc.x2 = rc.right;
drc.y1 = rc.top;
drc.y2 = rc.bottom;
//D3D::dev->Clear(1, &drc, D3DCLEAR_STENCIL|D3DCLEAR_TARGET|D3DCLEAR_ZBUFFER,clearColor,1.0f,0);
//if ((clearColor>>24) == 255)
D3D::dev->ColorFill(D3D::GetBackBufferSurface(), &rc, clearColor);
}
else
{
// TODO:
// bpmem.blendmode.alphaupdate
// bpmem.blendmode.colorupdate
// i dunno how to implement a clear on alpha only or color only
}
// clear z-buffer
if (bpmem.zmode.updateenable)
{
float clearZ = (float)bpmem.clearZValue / float(0xFFFFFF);
if (clearZ > 1.0f) clearZ = 1.0f;
if (clearZ < 0.0f) clearZ = 0.0f;
D3D::dev->Clear(0, 0, D3DCLEAR_ZBUFFER|D3DCLEAR_STENCIL, 0, clearZ, 0);
}
}
}
break;
case BPMEM_LOADTLUT: //GXLoadTlut
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
u32 tlutTMemAddr = (newval & 0x3FF) << 9;
u32 tlutXferCount = (newval & 0x1FFC00) >> 5;
u8 *ptr = 0;
// TODO - figure out a cleaner way.
if (g_VideoInitialize.bWii)
ptr = g_VideoInitialize.pGetMemoryPointer(bpmem.tlutXferSrc << 5);
else
ptr = g_VideoInitialize.pGetMemoryPointer((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 0xf6: // ksel0
case 0xf7: // ksel1
case 0xf8: // ksel2
@ -475,12 +570,12 @@ void BPWritten(int addr, int changes, int newval)
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
// PixelShaderMngr::SetTevKSelChanged(addr-0xf6);
PixelShaderManager::SetTevKSelChanged(addr-0xf6);
}
break;
default:
switch(addr & 0xF8) //texture sampler filter
switch (addr & 0xF8) //texture sampler filter
{
case 0x80: // TEX MODE 0
case 0xA0:
@ -526,282 +621,112 @@ void BPWritten(int addr, int changes, int newval)
//g_VideoInitialize.pLog(temp);
}
break;
case 0x84://TEX MODE 1
case 0xA4:
break;
case 0x88://TEX IMAGE 0
case 0xA8:
if (changes)
{
textureChanged[((addr&0xE0)==0xA0)*4+(addr&3)] = true;
VertexManager::Flush();
}
break;
case 0x8C://TEX IMAGE 1
case 0xAC:
if (changes)
{
textureChanged[((addr&0xE0)==0xA0)*4+(addr&3)] = true;
VertexManager::Flush();
}
break;
case 0x90://TEX IMAGE 2
case 0xB0:
if (changes)
{
textureChanged[((addr&0xE0)==0xA0)*4+(addr&3)] = true;
VertexManager::Flush();
}
break;
case 0x94://TEX IMAGE 3
case 0xB4:
if (changes)
{
textureChanged[((addr&0xE0)==0xA0)*4+(addr&3)] = true;
VertexManager::Flush();
}
break;
case 0x98://TEX TLUT
case 0xB8:
if (changes)
{
textureChanged[((addr&0xE0)==0xA0)*4+(addr&3)] = true;
VertexManager::Flush();
}
break;
case 0x9C://TEX UNKNOWN
case 0xBC:
break;
case 0x84://TEX MODE 1
case 0xA4:
break;
case 0x88://TEX IMAGE 0
case 0xA8:
if (changes)
{
//textureChanged[((addr&0xE0)==0xA0)*4+(addr&3)] = true;
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
}
break;
case 0x8C://TEX IMAGE 1
case 0xAC:
if (changes)
{
//textureChanged[((addr&0xE0)==0xA0)*4+(addr&3)] = true;
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
}
break;
case 0x90://TEX IMAGE 2
case 0xB0:
if (changes)
{
//textureChanged[((addr&0xE0)==0xA0)*4+(addr&3)] = true;
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
}
break;
case 0x94://TEX IMAGE 3
case 0xB4:
if (changes)
{
//textureChanged[((addr&0xE0)==0xA0)*4+(addr&3)] = true;
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
}
break;
case 0x98://TEX TLUT
case 0xB8:
if (changes)
{
//textureChanged[((addr&0xE0)==0xA0)*4+(addr&3)] = true;
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
}
break;
case 0x9C://TEX UNKNOWN
case 0xBC:
//ERROR_LOG("texunknown%x = %x\n", addr, newval);
((u32*)&bpmem)[addr] = newval;
break;
case 0xE0:
case 0xE4:
((u32*)&bpmem)[addr] = newval;
if (addr&1) { // don't compare with changes!
VertexManager::Flush();
int num = (addr >> 1) & 0x3;
PixelShaderManager::SetColorChanged(bpmem.tevregs[num].high.type, num);
}
break;
default:
switch(addr&0xF0) {
case 0x30:
{
int tc = addr&0x1;
int stage = (addr>>1)&0x7;
TCoordInfo &tci = bpmem.texcoords[stage];
//TCInfo &t = (tc?tci.s:tc.t);
// cylindric wrapping here
//dev->SetRenderState(D3DRS_WRAP0+stage, D3DWRAPCOORD_0);
}
break;
switch (addr&0xF0)
{
case 0x10: // tevindirect 0-15
if (changes) {
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
PixelShaderManager::SetTevIndirectChanged(addr - 0x10);
}
break;
case 0x30:
if (changes) {
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
PixelShaderManager::SetTexDimsChanged((addr >> 1) & 0x7);
}
break;
case 0xC0:
case 0xD0:
if (changes)
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
// PixelShaderMngr::SetTevCombinerChanged((addr&0x1f)/2);
PixelShaderManager::SetTevCombinerChanged((addr & 0x1f)/2);
}
break;
case 0xE0:
if (addr<0xe8)
{
if (addr&1)
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
static int lastRGBA[2][4] = {
{0xEEEEEEEE, 0xEEEEEEEE, 0xEEEEEEEE, 0xEEEEEEEE},
{0xEEEEEEEE, 0xEEEEEEEE, 0xEEEEEEEE, 0xEEEEEEEE}
};
//Terrible hack
//The reason is that there are two sets of registers
//overloaded here...
int num = (addr >> 1) & 0x3;
int type = bpmem.tevregs[num].high.type;
int colorbase = type ? PS_CONST_KCOLORS : PS_CONST_COLORS;
int r=bpmem.tevregs[num].low.a, a=bpmem.tevregs[num].low.b;
int b=bpmem.tevregs[num].high.a, g=bpmem.tevregs[num].high.b;
int rgba = ((a<<24) | (r << 16) | (g << 8) | b) & 0xFCFCFCFC; //let's not detect minimal changes
if (rgba != lastRGBA[type][num])
{
VertexManager::Flush();
lastRGBA[type][num] = rgba;
float temp[4] = {
r/255.0f, g/255.0f, b/255.0f, a/255.0f
};
D3D::dev->SetPixelShaderConstantF(colorbase + num, temp, 1);
}
}
}
break;
case 0x20:
case 0x80:
case 0x90:
case 0xA0:
case 0xB0:
default:
if (changes)
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
}
case 0x20:
case 0x80:
case 0x90:
case 0xA0:
case 0xB0:
default:
break;
}
break;
}
break;
}
}
// __________________________________________________________________________________________________
// LoadBPReg
//
void LoadBPReg(u32 value0)
{
DVSTARTPROFILE();
//handle the mask register
int opcode = value0 >> 24;
int oldval = ((u32*)&bpmem)[opcode];
int newval = (((u32*)&bpmem)[opcode] & ~bpmem.bpMask) | (value0 & bpmem.bpMask);
int changes = (oldval ^ newval) & 0xFFFFFF;
//reset the mask register
if(opcode != 0xFE)
bpmem.bpMask = 0xFFFFFF;
switch (opcode)
{
case 0x45: //GXSetDrawDone
VertexManager::Flush();
switch (value0 & 0xFF)
{
case 0x02:
g_VideoInitialize.pSetPEFinish(); // may generate interrupt
DebugLog("GXSetDrawDone SetPEFinish (value: 0x%02X)", (value0 & 0xFFFF));
break;
default:
DebugLog("GXSetDrawDone ??? (value 0x%02X)", (value0 & 0xFFFF));
break;
}
break;
case BPMEM_PE_TOKEN_ID:
g_VideoInitialize.pSetPEToken(static_cast<u16>(value0 & 0xFFFF), FALSE);
DebugLog("SetPEToken 0x%04x", (value0 & 0xFFFF));
break;
case BPMEM_PE_TOKEN_INT_ID:
g_VideoInitialize.pSetPEToken(static_cast<u16>(value0 & 0xFFFF), TRUE);
DebugLog("SetPEToken + INT 0x%04x", (value0 & 0xFFFF));
break;
case 0x67: // set gp metric?
break;
case 0x52:
{
VertexManager::Flush();
((u32*)&bpmem)[opcode] = newval;
RECT rc = {
(LONG)(bpmem.copyTexSrcXY.x*Renderer::GetXScale()),
(LONG)(bpmem.copyTexSrcXY.y*Renderer::GetYScale()),
(LONG)((bpmem.copyTexSrcXY.x+bpmem.copyTexSrcWH.x)*Renderer::GetXScale()),
(LONG)((bpmem.copyTexSrcXY.y+bpmem.copyTexSrcWH.y)*Renderer::GetYScale())
};
UPE_Copy PE_copy;
PE_copy.Hex = bpmem.triggerEFBCopy;
// clamp0
// clamp1
// target_pixel_format
// gamma
// scale_something
// clear
// frame_to_field
// copy_to_xfb
// ???: start Mem to/from EFB transfer
/* bool bMip = false; // ignored
if (bpmem.triggerEFBCopy & EFBCOPY_GENERATEMIPS)
bMip = true;*/
if (PE_copy.copy_to_xfb == 0) // bpmem.triggerEFBCopy & EFBCOPY_EFBTOTEXTURE)
{
// EFB to texture
// for some reason it sets bpmem.zcontrol.pixel_format to PIXELFMT_Z24 every time a zbuffer format is given as a dest to GXSetTexCopyDst
TextureCache::CopyEFBToRenderTarget(bpmem.copyTexDest<<5, &rc);
}
else
{
// EFB to XFB
// MessageBox(0, "WASDF", 0, 0);
Renderer::SwapBuffers();
DebugLog("Renderer::SwapBuffers()");
g_VideoInitialize.pCopiedToXFB();
}
// clearing
if (PE_copy.clear) // bpmem.triggerEFBCopy & EFBCOPY_CLEAR)
{
// 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
// clear color
u32 clearColor = (bpmem.clearcolorAR<<16)|bpmem.clearcolorGB;
if (bpmem.blendmode.colorupdate)
{
D3DRECT drc;
drc.x1 = rc.left;
drc.x2 = rc.right;
drc.y1 = rc.top;
drc.y2 = rc.bottom;
//D3D::dev->Clear(1, &drc, D3DCLEAR_STENCIL|D3DCLEAR_TARGET|D3DCLEAR_ZBUFFER,clearColor,1.0f,0);
//if ((clearColor>>24) == 255)
D3D::dev->ColorFill(D3D::GetBackBufferSurface(), &rc, clearColor);
}
else
{
// TODO:
// bpmem.blendmode.alphaupdate
// bpmem.blendmode.colorupdate
// i dunno how to implement a clear on alpha only or color only
}
// clear z-buffer
if (bpmem.zmode.updateenable)
{
float clearZ = (float)bpmem.clearZValue / float(0xFFFFFF);
if (clearZ > 1.0f) clearZ = 1.0f;
if (clearZ < 0.0f) clearZ = 0.0f;
D3D::dev->Clear(0, 0, D3DCLEAR_ZBUFFER|D3DCLEAR_STENCIL, 0, clearZ, 0);
}
}
}
break;
case 0x65: //GXLoadTlut
{
VertexManager::Flush();
((u32*)&bpmem)[opcode] = newval;
u32 tlutTMemAddr = (value0&0x3FF)<<9;
u32 tlutXferCount = (value0&0x1FFC00)>>5;
//do the transfer!!
memcpy(texMem + tlutTMemAddr, g_VideoInitialize.pGetMemoryPointer((bpmem.tlutXferSrc&0xFFFFF)<<5), tlutXferCount);
// 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;
}
//notify the video handling so it can update render states
BPWritten(opcode, changes, newval);
((u32*)&bpmem)[opcode] = newval;
}
void BPReload()
{
for (int i=0; i<254; i++)
BPWritten(i, 0xFFFFFF, ((u32*)&bpmem)[i]);
}
void ActivateTextures()
{
for (int i = 0; i < 8; i++)
@ -821,5 +746,3 @@ void ActivateTextures()
textureChanged[i] = false;
}
}

3
Source/Plugins/Plugin_VideoDX9/Src/CPStructs.cpp
View File

@ -19,7 +19,6 @@
#include "CPStructs.h"
#include "XFStructs.h"
#include "TransformEngine.h"
#include "VertexManager.h"
#include "VertexLoader.h"
@ -27,7 +26,7 @@
void CPUpdateMatricesA()
{
float *flipmem = (float *)xfmem;
const float *flipmem = (const float *)xfmem;
CTransformEngine::SetPosNormalMatrix(
flipmem + MatrixIndexA.PosNormalMtxIdx * 4, //CHECK
flipmem + 0x400 + 3 * (MatrixIndexA.PosNormalMtxIdx & 31)); //CHECK

19
Source/Plugins/Plugin_VideoDX9/Src/D3DShader.cpp
View File

@ -24,7 +24,7 @@
namespace D3D
{
LPDIRECT3DVERTEXSHADER9 CompileVShader(const char *code, int len)
LPDIRECT3DVERTEXSHADER9 CompileVertexShader(const char *code, int len)
{
//try to compile
LPD3DXBUFFER shaderBuffer = 0;
@ -43,8 +43,7 @@ LPDIRECT3DVERTEXSHADER9 CompileVShader(const char *code, int len)
std::string hello = (char*)errorBuffer->GetBufferPointer();
hello += "\n\n";
hello += code;
if (g_Config.bShowShaderErrors)
MessageBox(0, hello.c_str(), "Error compiling vertex shader", MB_ICONERROR);
MessageBox(0, hello.c_str(), "Error compiling vertex shader", MB_ICONERROR);
vShader = 0;
}
else if (SUCCEEDED(hr))
@ -55,8 +54,7 @@ LPDIRECT3DVERTEXSHADER9 CompileVShader(const char *code, int len)
hr = D3D::dev->CreateVertexShader((DWORD *)shaderBuffer->GetBufferPointer(), &vShader);
if (FAILED(hr) || vShader == 0)
{
if (g_Config.bShowShaderErrors)
MessageBox(0,code,(char*)errorBuffer->GetBufferPointer(),MB_ICONERROR);
MessageBox(0, code, (char*)errorBuffer->GetBufferPointer(),MB_ICONERROR);
}
}
@ -69,7 +67,7 @@ LPDIRECT3DVERTEXSHADER9 CompileVShader(const char *code, int len)
return vShader;
}
LPDIRECT3DPIXELSHADER9 CompilePShader(const char *code, int len)
LPDIRECT3DPIXELSHADER9 CompilePixelShader(const char *code, int len)
{
LPD3DXBUFFER shaderBuffer = 0;
LPD3DXBUFFER errorBuffer = 0;
@ -83,8 +81,7 @@ LPDIRECT3DPIXELSHADER9 CompilePShader(const char *code, int len)
std::string hello = (char*)errorBuffer->GetBufferPointer();
hello += "\n\n";
hello += code;
if (g_Config.bShowShaderErrors)
MessageBox(0, hello.c_str(), "Error compiling pixel shader", MB_ICONERROR);
MessageBox(0, hello.c_str(), "Error compiling pixel shader", MB_ICONERROR);
pShader = 0;
}
else
@ -93,8 +90,7 @@ LPDIRECT3DPIXELSHADER9 CompilePShader(const char *code, int len)
HRESULT hr = D3D::dev->CreatePixelShader((DWORD *)shaderBuffer->GetBufferPointer(), &pShader);
if (FAILED(hr) || pShader == 0)
{
if (g_Config.bShowShaderErrors)
MessageBox(0,"damn","error creating pixelshader",MB_ICONERROR);
MessageBox(0, "damn", "error creating pixelshader", MB_ICONERROR);
}
}
@ -106,4 +102,5 @@ LPDIRECT3DPIXELSHADER9 CompilePShader(const char *code, int len)
return pShader;
}
}
} // namespace

4
Source/Plugins/Plugin_VideoDX9/Src/D3DShader.h
View File

@ -21,6 +21,6 @@
namespace D3D
{
LPDIRECT3DVERTEXSHADER9 CompileVShader(const char *code, int len);
LPDIRECT3DPIXELSHADER9 CompilePShader(const char *code, int len);
LPDIRECT3DVERTEXSHADER9 CompileVertexShader(const char *code, int len);
LPDIRECT3DPIXELSHADER9 CompilePixelShader(const char *code, int len);
}

2
Source/Plugins/Plugin_VideoDX9/Src/D3DUtil.cpp
View File

@ -83,7 +83,7 @@ namespace D3D
HFONT hFont = CreateFont(nHeight, 0, 0, 0, dwBold, 0,
FALSE, FALSE, DEFAULT_CHARSET, OUT_DEFAULT_PRECIS,
CLIP_DEFAULT_PRECIS, ANTIALIASED_QUALITY,
VARIABLE_PITCH, "Courier New");
VARIABLE_PITCH, "Tahoma");
if (NULL == hFont)
return E_FAIL;

92
Source/Plugins/Plugin_VideoDX9/Src/DecodedVArray.cpp
View File

@ -1,92 +0,0 @@
// 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 "stdafx.h"
#include "DecodedVArray.h"
#include "main.h"
DecodedVArray::DecodedVArray()
{
Zero();
}
DecodedVArray::~DecodedVArray()
{
Destroy();
}
void DecodedVArray::Zero()
{
size = 0;
count = 0;
components = 0;
positions = 0;
posMtxInds = 0;
for (int i=0; i<3; i++)
normals[i] = 0;
for (int i=0; i<2; i++)
colors[i] = 0;
for (int i=0; i<8; i++)
{
texMtxInds[i] = 0;
uvs[i] = 0;
}
}
void DecodedVArray::Destroy()
{
//,,
delete [] positions;
delete [] posMtxInds;
for (int i=0; i<3; i++)
delete [] normals[i];
for (int i=0; i<2; i++)
delete [] colors[i];
for (int i=0; i<8; i++)
{
delete [] uvs[i];
delete [] texMtxInds[i];
}
Zero();
}
void DecodedVArray::Create(int _size, int pmcount, int tmcount, int nrmcount, int colcount, int tccount)
{
size = _size;
// position matrix indices
if (pmcount)
posMtxInds = new DecMtxInd[size];
// texture matrix indices
if (tmcount)
for (int i=0; i<tmcount; i++)
texMtxInds[i] = new DecMtxInd[size];
// positions (always)
positions = new DecPos[size];
// normals
if (nrmcount)
for (int i=0; i<nrmcount; i++)
normals[i] = new DecNormal[size];
// colors
if (colcount)
for (int i=0; i<colcount; i++)
colors[i] = new DecColor[size];
if (tccount)
for (int i=0; i<tccount; i++)
uvs[i] = new DecUV[size];
}

114
Source/Plugins/Plugin_VideoDX9/Src/DecodedVArray.h
View File

@ -1,114 +0,0 @@
// 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/
#ifndef _DECODED_VARRAY_H
#define _DECODED_VARRAY_H
#include "Vec3.h"
#include "Common.h"
typedef Vec3 DecPos;
typedef Vec3 DecNormal;
struct DecUV
{
float u,v;
};
typedef u32 DecColor;
typedef u8 DecMtxInd;
int ComputeVertexSize(u32 components);
//TODO(ector): Change the internal implementation to pack it tight according to components
// The tight packing will be fed directly to the gfx card in the mystic future.
class DecodedVArray
{
int size;
u32 components;
int vertexSize;
public:
int count;
DecodedVArray();
~DecodedVArray();
void SetComponents(u32 comps) {components = comps; vertexSize = ComputeVertexSize(components);
ComputeComponents(); }
u32 GetComponents() const {return components;}
void Create(int _size, int pmcount, int tmcount, int nrmcount, int colcount, int tccount);
void Zero();
void Destroy();
void Reset() {count=0;}
int GetSize() {return size;}
int GetCount() {return count;}
void Next() {count++;}
void SetPosNrmIdx(int i) {posMtxInds[count] = i;}
void SetTcIdx(int n, int i) {texMtxInds[n][count] = i;}
void SetPosX(float x) {positions[count].x=x;}
void SetPosY(float y) {positions[count].y=y;}
void SetPosZ(float z) {positions[count].z=z;}
void SetNormalX(int n,float x) {normals[n][count].x=x;}
void SetNormalY(int n,float y) {normals[n][count].y=y;}
void SetNormalZ(int n,float z) {normals[n][count].z=z;}
void SetU(int n, float u) {uvs[n][count].u = u;}
void SetV(int n, float v) {uvs[n][count].v = v;}
void SetPosition(float x, float y, float z) {
positions[count].x=x;
positions[count].y=y;
positions[count].z=z;
}
void SetNormal(int n, float x, float y, float z) {
normals[n][count].x=x;
normals[n][count].y=y;
normals[n][count].z=z;
}
void SetColor(int n, u32 c)
{
colors[n][count] = c;
}
void SetUV(int n, float u, float v) {
uvs[n][count].u=u;
uvs[n][count].v=v;
}
void ComputeComponents() {
hasPosMatIdx = (components & (1 << 1)) != 0;
for(int i = 0; i < 8; i++)
hasTexMatIdx[i] = (components & (1 << (i + 2))) != 0;
hasNrm = (components & (1 << 10)) != 0;
}
const DecPos &GetPos(int n) const { return positions[n]; }
const DecNormal &GetNormal(int i, int n) const { return normals[i][n]; }
const DecColor &GetColor(int i, int n) const { return colors[i][n]; }
const DecUV &GetUV(int i, int n) const { return uvs[i][n]; }
const DecMtxInd &GetPosMtxInd(int n) const { return posMtxInds[n]; }
const DecMtxInd &GetTexMtxInd(int i, int n) const { return texMtxInds[i][n]; }
//private:
DecPos *positions;
DecNormal *normals[3];
DecColor *colors[2];
DecUV *uvs[8];
DecMtxInd *posMtxInds;
DecMtxInd *texMtxInds[8];
// Component data
bool hasPosMatIdx, hasTexMatIdx[8], hasNrm;
};
#endif

1
Source/Plugins/Plugin_VideoDX9/Src/EmuWindow.cpp
View File

@ -52,6 +52,7 @@ namespace EmuWindow
break;
case WM_CLOSE:
exit(0);
return 0;
case WM_DESTROY:

16
Source/Plugins/Plugin_VideoDX9/Src/NativeVertexFormat.cpp
View File

@ -1,3 +1,4 @@
// Copyright (C) 2003-2008 Dolphin Project.
// This program is free software: you can redistribute it and/or modify
@ -21,7 +22,7 @@
#include "x64Emitter.h"
#include "ABI.h"
#include "MemoryUtil.h"
#include "VertexShader.h"
#include "VertexShaderGen.h"
#include "CPMemory.h"
#include "NativeVertexFormat.h"
@ -29,7 +30,6 @@
class D3DVertexFormat : public NativeVertexFormat
{
PortableVertexDeclaration vtx_decl;
LPDIRECT3DVERTEXDECLARATION9 d3d_decl;
public:
@ -39,7 +39,6 @@ public:
virtual void SetupVertexPointers() const;
};
NativeVertexFormat *NativeVertexFormat::Create()
{
return new D3DVertexFormat();
@ -58,7 +57,6 @@ D3DVertexFormat::~D3DVertexFormat()
}
}
D3DDECLTYPE VarToD3D(VarType t)
{
static const D3DDECLTYPE lookup[5] =
@ -74,6 +72,8 @@ D3DDECLTYPE VarToD3D(VarType t)
void D3DVertexFormat::Initialize(const PortableVertexDeclaration &_vtx_decl)
{
vertex_stride = _vtx_decl.stride;
D3DVERTEXELEMENT9 *elems = new D3DVERTEXELEMENT9[32];
memset(elems, 0, sizeof(D3DVERTEXELEMENT9) * 32);
@ -123,8 +123,9 @@ void D3DVertexFormat::Initialize(const PortableVertexDeclaration &_vtx_decl)
}
}
if (vtx_decl.posmtx_offset != -1)
if (_vtx_decl.posmtx_offset != -1)
{
PanicAlert("boo %i", _vtx_decl.posmtx_offset);
// glVertexAttribPointer(SHADER_POSMTX_ATTRIB, 4, GL_UNSIGNED_BYTE, GL_FALSE, vtx_decl.stride, (void *)vtx_decl.posmtx_offset);
elems[elem_idx].Offset = _vtx_decl.posmtx_offset;
elems[elem_idx].Usage = D3DDECLUSAGE_BLENDINDICES;
@ -132,6 +133,11 @@ void D3DVertexFormat::Initialize(const PortableVertexDeclaration &_vtx_decl)
++elem_idx;
}
// End marker
elems[elem_idx].Stream = 0xff;
elems[elem_idx].Type = D3DDECLTYPE_UNUSED;
++elem_idx;
if (FAILED(D3D::dev->CreateVertexDeclaration(elems, &d3d_decl)))
{
PanicAlert("Failed to create D3D vertex declaration!");

525
Source/Plugins/Plugin_VideoDX9/Src/PixelShader.cpp
View File

@ -1,525 +0,0 @@
// 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 "PixelShader.h"
#include "BPStructs.h"
#include "XFStructs.h"
/*
old tev->pixelshader notes
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
*/
const float epsilon = 1.0f/255.0f;
const char *tevKSelTableC[] =
{
"1,1,1", //KCSEL_1 = 0x00
"0.875,0.875,0.875",//KCSEL_7_8 = 0x01
"0.75,0.75,0.75", //KCSEL_3_4 = 0x02
"0.625,0.625,0.625",//KCSEL_5_8 = 0x03
"0.5,0.5,0.5", //KCSEL_1_2 = 0x04
"0.375,0.375,0.375",//KCSEL_3_8 = 0x05
"0.25,0.25,0.25", //KCSEL_1_4 = 0x06
"0.125,0.125,0.125",//KCSEL_1_8 = 0x07
"ERROR", //0x08
"ERROR", //0x09
"ERROR", //0x0a
"ERROR", //0x0b
"k0.rgb",//KCSEL_K0 = 0x0C
"k1.rgb",//KCSEL_K1 = 0x0D
"k2.rgb",//KCSEL_K2 = 0x0E
"k3.rgb",//KCSEL_K3 = 0x0F
"k0.rrr",//KCSEL_K0_R = 0x10
"k1.rrr",//KCSEL_K1_R = 0x11
"k2.rrr",//KCSEL_K2_R = 0x12
"k3.rrr",//KCSEL_K3_R = 0x13
"k0.ggg",//KCSEL_K0_G = 0x14
"k1.ggg",//KCSEL_K1_G = 0x15
"k2.ggg",//KCSEL_K2_G = 0x16
"k3.ggg",//KCSEL_K3_G = 0x17
"k0.bbb",//KCSEL_K0_B = 0x18
"k1.bbb",//KCSEL_K1_B = 0x19
"k2.bbb",//KCSEL_K2_B = 0x1A
"k3.bbb",//KCSEL_K3_B = 0x1B
"k0.aaa",//KCSEL_K0_A = 0x1C
"k1.aaa",//KCSEL_K1_A = 0x1D
"k2.aaa",//KCSEL_K2_A = 0x1E
"k3.aaa",//KCSEL_K3_A = 0x1F
};
const char *tevKSelTableA[] =
{
"1", //KASEL_1 = 0x00
"0.875",//KASEL_7_8 = 0x01
"0.75", //KASEL_3_4 = 0x02
"0.625",//KASEL_5_8 = 0x03
"0.5", //KASEL_1_2 = 0x04
"0.375",//KASEL_3_8 = 0x05
"0.25", //KASEL_1_4 = 0x06
"0.125",//KASEL_1_8 = 0x07
"ERROR",//0x08
"ERROR",//0x09
"ERROR",//0x0a
"ERROR",//0x0b
"ERROR",//0x0c
"ERROR",//0x0d
"ERROR",//0x0e
"ERROR",//0x0f
"k0.r", //KASEL_K0_R = 0x10
"k1.r", //KASEL_K1_R = 0x11
"k2.r", //KASEL_K2_R = 0x12
"k3.r", //KASEL_K3_R = 0x13
"k0.g", //KASEL_K0_G = 0x14
"k1.g", //KASEL_K1_G = 0x15
"k2.g", //KASEL_K2_G = 0x16
"k3.g", //KASEL_K3_G = 0x17
"k0.b", //KASEL_K0_B = 0x18
"k1.b", //KASEL_K1_B = 0x19
"k2.b", //KASEL_K2_B = 0x1A
"k3.b", //KASEL_K3_B = 0x1B
"k0.a", //KASEL_K0_A = 0x1C
"k1.a", //KASEL_K1_A = 0x1D
"k2.a", //KASEL_K2_A = 0x1E
"k3.a", //KASEL_K3_A = 0x1F
};
const char *tevScaleTable[] =
{
"1", //SCALE_1
"2", //SCALE_2
"4", //SCALE_4
"0.5", //DIVIDE_2
};
const char *tevBiasTable[] =
{
"", //ZERO,
"+0.5", //ADD_HALF,
"-0.5", //SUB_HALF,
"", //WTF? seen in shadow2
};
const char *tevOpTable[] =
{
"+", //ADD = 0,
"-", //SUB = 1,
};
const char *tevCompOpTable[] =
{
">",
"==",
};
#define TEV_COMP_R8 0
#define TEV_COMP_GR16 1
#define TEV_COMP_BGR24 2
#define TEV_COMP_RGB8 3
const char *tevCInputTable[] =
{
"prev.rgb", //CPREV,
"prev.aaa", //APREV,
"c0.rgb", //C0,
"c0.aaa", //A0,
"c1.rgb", //C1,
"c1.aaa", //A1,
"c2.rgb", //C2,
"c2.aaa", //A2,
"textemp.rgb", //TEXC,
"textemp.aaa", //TEXA,
"rastemp.rgb", //RASC,
"rastemp.aaa", //RASA,
"float3(1,1,1)", //ONE,
"float3(.5,.5,.5)", //HALF,
"konsttemp.rgb", //KONST,
"float3(0,0,0)", //ZERO
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
};
const char *tevCInputTable2[] =
{
"prev", //CPREV,
"(prev.aaa)", //APREV,
"c0", //C0,
"(c0.aaa)", //A0,
"c1", //C1,
"(c1.aaa)", //A1,
"c2", //C2,
"(c2.aaa)", //A2,
"textemp", //TEXC,
"(textemp.aaa)", //TEXA,
"rastemp", //RASC,
"(rastemp.aaa)", //RASA,
"float3(1,1,1)", //ONE,
"float3(.5,.5,.5)", //HALF,
"konsttemp", //KONST,
"float3(0,0,0)", //ZERO
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
};
const char *tevAInputTable[] =
{
"prev.a", //APREV,
"c0.a", //A0,
"c1.a", //A1,
"c2.a", //A2,
"textemp.a", //TEXA,
"rastemp.a", //RASA,
"konsttemp.a", //KONST, (hw1 had quarter)
"0.0", //ZERO
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
};
const char *tevAInputTable1[] =
{
"prev.r", //APREV,
"c0.r", //A0,
"c1.r", //A1,
"c2.r", //A2,
"textemp.r", //TEXA,
"rastemp.r", //RASA,
"konsttemp.r", //KONST, (hw1 had quarter)
"0.0", //ZERO
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
};
const char *tevAInputTable2[] =
{
"prev", //APREV,
"c0", //A0,
"c1", //A1,
"c2", //A2,
"textemp", //TEXA,
"rastemp", //RASA,
"konsttemp", //KONST, (hw1 had quarter)
"float4(0,0,0,0)",//ZERO
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
"PADERROR", "PADERROR", "PADERROR", "PADERROR",
};
const char *tevRasTable[] =
{
"colors[0]",//RAS1_CC_0 0x00000000 /* color channel 0 */
"colors[1]",//RAS1_CC_1 0x00000001 /* color channel 1 */
"ERROR", //2
"ERROR", //3
"ERROR", //4
"alphabump", //RAS1_CC_B 0x00000005 /* indirect texture bump alpha */ //green cuz unsupported
"(alphabump*(255.0f/248.0f))", //RAS1_CC_BN 0x00000006 /* ind tex bump alpha, normalized 0-255 *///green cuz unsupported
"float4(0,0,0,0)", //RAS1_CC_Z 0x00000007 /* set color value to zero */
};
const char *tevCOutputTable[] = { "prev.rgb", "c0.rgb", "c1.rgb", "c2.rgb" };
const char *tevAOutputTable[] = { "prev.a", "c0.a", "c1.a", "c2.a" };
const char *tevIndAlphaSel[] = {"", "x", "y", "z"};
const char *tevIndAlphaScale[] = {"", "*32","*16","*8"};
const char *tevIndBiasField[] = {"", "x", "y", "xy", "z", "xz", "yz", "xyz"}; // indexed by bias
const char *tevIndBiasAdd[] = {"-128.0f", "1.0f", "1.0f", "1.0f" }; // indexed by fmt
const char *tevIndWrapStart[] = {"0", "256", "128", "64", "32", "16", "0.001" };
const char *tevIndFmtScale[] = {"255.0f", "31.0f", "15.0f", "8.0f" };
const char *tevTexFuncs[] =
{
"tex2D",
"tex2Dproj"
};
const char *alphaRef[2] =
{
"alphaRef.x",
"alphaRef.y"
};
char text[65536];
#define WRITE p+=sprintf
void WriteStage(char *&p, int n);
void WriteAlphaTest(char *&p);
char *swapColors = "rgba";
char swapModeTable[4][5];
void BuildSwapModeTable()
{
for (int i = 0; i < 4; i++)
{
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;
}
}
const char *GeneratePixelShader()
{
BuildSwapModeTable();
int numStages = bpmem.genMode.numtevstages + 1;
int numTexgen = bpmem.genMode.numtexgens;
int numSamplers = 8;
char *p = text;
WRITE(p,"//Pixel Shader for TEV stages\n\
//%i TEV stages, %i texgens, %i IND stages, %i COL channels\n",
bpmem.genMode.numtevstages,bpmem.genMode.numtexgens,bpmem.genMode.numindstages,bpmem.genMode.numcolchans);
//write kcolor declarations
for (int i = 0; i < 4; i++) {
if(i < 3) {
WRITE(p,"float4 k%i : register(c%i);\n\
float4 color%i : register(c%i);\n",i,PS_CONST_KCOLORS+i, i,PS_CONST_COLORS+i+1);
} else {
WRITE(p,"float4 k%i : register(c%i);\n",i,PS_CONST_KCOLORS+i);
}
}
WRITE(p,"float constalpha : register(c%i);\n\
float2 alphaRef : register(c%i);\n\n\
sampler samp[%i] : register(s0);\n\n\
float4 main(in float4 colors[2] : COLOR0",PS_CONST_CONSTALPHA,PS_CONST_ALPHAREF,numSamplers);
if (numTexgen)
WRITE(p,", float4 uv[%i] : TEXCOORD0",numTexgen);
else
WRITE(p,", float4 uv[1] : TEXCOORD0"); //HACK
WRITE(p,") : COLOR\n\
{\n\
float4 c0=color0,c1=color1,c2=color2,prev=float4(0.0f,0.0f,0.0f,0.0f),textemp,rastemp,konsttemp;\n\
float3 comp16 = float3(1,255,0), comp24 = float3(1,255,255*255);\n\
\n");
for (int i = 0; i < numStages; i++)
WriteStage(p,i); //build the equation for this stage
WriteAlphaTest(p);
/* see GL shader generator - this is Donko's hack
if (bpmem.dstalpha.enable)
WRITE(p," return float4(prev.rgb,constalpha.x);\n");
else
*/
WRITE(p," return prev;\n");
WRITE(p,"}\n\0");
return text;
}
void WriteStage(char *&p, int n)
{
const char *rasswap = swapModeTable[bpmem.combiners[n].alphaC.rswap];
const char *texswap = swapModeTable[bpmem.combiners[n].alphaC.tswap];
int texfun = xfregs.texcoords[n].texmtxinfo.projection;
WRITE(p,"rastemp=%s.%s;\n",tevRasTable[bpmem.tevorders[n/2].getColorChan(n&1)],rasswap);
if (bpmem.tevorders[n/2].getEnable(n&1))
WRITE(p,"textemp=%s(samp[%i],uv[%i]).%s;\n",
tevTexFuncs[texfun],
bpmem.tevorders[n/2].getTexMap(n&1),
bpmem.tevorders[n/2].getTexCoord(n&1),texswap);
else
WRITE(p,"textemp=float4(1,1,1,1);\n");
int kc = bpmem.tevksel[n/2].getKC(n&1);
int ka = bpmem.tevksel[n/2].getKA(n&1);
WRITE(p,"konsttemp=float4(%s,%s);\n",tevKSelTableC[kc],tevKSelTableA[ka]);
TevStageCombiner::ColorCombiner &cc = bpmem.combiners[n].colorC;
TevStageCombiner::AlphaCombiner &ac = bpmem.combiners[n].alphaC;
WRITE(p,"float4(%s,%s)=", tevCOutputTable[cc.dest], tevAOutputTable[ac.dest]);
//////////////////////////////////////////////////////////////////////////
//start of color
//////////////////////////////////////////////////////////////////////////
WRITE(p,"float4(\n");
if (cc.bias != TB_COMPARE)
{
//normal color combiner goes here
WRITE(p," %s*(%s%s",tevScaleTable[cc.shift],tevCInputTable[cc.d],tevOpTable[cc.op]);
WRITE(p,"(lerp(%s,%s,%s)%s)),\n",
tevCInputTable[cc.a],tevCInputTable[cc.b],
tevCInputTable[cc.c],tevBiasTable[cc.bias]);
}
else
{
//compare color combiner goes here
switch(cc.shift) // yep comparemode stored here :P
{
case TEV_COMP_R8:
if (cc.op == 0) //equality check needs tolerance, fp in gpu has drawbacks :(
WRITE(p," %s + ((%s.r > %s.r) ? %s : float3(0,0,0)),\n",
tevCInputTable[cc.d],tevCInputTable2[cc.a],
tevCInputTable2[cc.b],tevCInputTable[cc.c]);
else
WRITE(p," %s + (abs(%s.r - %s.r)<%f ? %s : float3(0,0,0)),\n",
tevCInputTable[cc.d],tevCInputTable2[cc.a],
tevCInputTable2[cc.b],epsilon,tevCInputTable[cc.c]);
break;
default:
WRITE(p,"float3(0,0,0),\n");
break;
}
}
//end of color
//////////////////////////////////////////////////////////////////////////
//start of alpha
//////////////////////////////////////////////////////////////////////////
if (ac.bias != TB_COMPARE)
{
//normal alpha combiner goes here
WRITE(p," %s*(%s%s",tevScaleTable[ac.shift],tevAInputTable[ac.d],tevOpTable[ac.op]);
WRITE(p,"lerp(%s,%s,%s) %s)\n",
tevAInputTable[ac.a],tevAInputTable[ac.b],
tevAInputTable[ac.c],tevBiasTable[ac.bias]);
}
else
{
int cmp = (ac.shift<<1)|ac.op|8; // comparemode stored here
const char **inputTable = NULL;
inputTable = (cmp == TEVCMP_R8_GT || cmp == TEVCMP_R8_EQ) ? tevAInputTable1 : tevAInputTable;
//compare alpha combiner goes here
switch(cmp) {
case TEVCMP_R8_GT:
case TEVCMP_A8_GT:
WRITE(p," %s + ((%s > %s) ? %s : 0)\n",
tevAInputTable[ac.d], inputTable[ac.a], inputTable[ac.b], tevAInputTable[ac.c]);
break;
case TEVCMP_R8_EQ:
case TEVCMP_A8_EQ:
WRITE(p," %s + (abs(%s - %s)<%f ? %s : 0)\n",
tevAInputTable[ac.d], inputTable[ac.a], inputTable[ac.b],epsilon,tevAInputTable[ac.c]);
break;
case TEVCMP_GR16_GT: // 16 bit compares: 255*g+r (probably used for ztextures, so make sure in ztextures, g is the most significant byte)
case TEVCMP_BGR24_GT: // 24 bit compares: 255*255*b+255*g+r
WRITE(p," %s + (( dot(%s.rgb-%s.rgb, comp%s) > 0) ? %s : 0)\n",
tevAInputTable[ac.d],tevAInputTable2[ac.a], tevAInputTable2[ac.b], cmp==TEVCMP_GR16_GT?"16":"24", tevAInputTable[ac.c]);
break;
case TEVCMP_GR16_EQ:
case TEVCMP_BGR24_EQ:
WRITE(p," %s + (abs(dot(%s.rgb - %s.rgb, comp%s))<%f ? %s : 0)\n",
tevAInputTable[ac.d],tevAInputTable2[ac.a], tevAInputTable2[ac.b],cmp==TEVCMP_GR16_EQ?"16":"24",epsilon,tevAInputTable[ac.c]);
break;
default:
WRITE(p,"0)\n");
break;
}
}
WRITE(p, ");");
if (ac.clamp)
WRITE(p, "%s = clamp(%s, 0.0f, 1.0f);\n", tevAOutputTable[ac.dest], tevAOutputTable[ac.dest]);
WRITE(p, "\n");
}
void WriteAlphaCompare(char *&p, int num, int comp)
{
WRITE(p," res%i = ",num);
switch(comp) {
case ALPHACMP_ALWAYS: WRITE(p,"0;\n"); break;
case ALPHACMP_NEVER: WRITE(p,"1;\n"); break;
case ALPHACMP_LEQUAL: WRITE(p,"prev.a - %s.x;\n",alphaRef[num]); break;
case ALPHACMP_LESS: WRITE(p,"prev.a - %s.x + %f;\n",alphaRef[num],epsilon*2);break;
case ALPHACMP_GEQUAL: WRITE(p,"%s - prev.a;\n",alphaRef[num]); break;
case ALPHACMP_GREATER: WRITE(p,"%s - prev.a + %f;\n",alphaRef[num],epsilon*2);break;
case ALPHACMP_EQUAL: WRITE(p,"abs(%s-prev.a)-%f;\n",alphaRef[num],epsilon*2); break;
case ALPHACMP_NEQUAL: WRITE(p,"%f-abs(%s-prev.a);\n",epsilon*2,alphaRef[num]); break;
}
}
void WriteAlphaTest(char *&p)
{
AlphaOp op = (AlphaOp)bpmem.alphaFunc.logic;
Compare comp[2] = {(Compare)bpmem.alphaFunc.comp0,(Compare)bpmem.alphaFunc.comp1};
//first kill all the simple cases
if (op == ALPHAOP_AND && (comp[0] == COMPARE_ALWAYS && comp[1] == COMPARE_ALWAYS)) return;
if (op == ALPHAOP_OR && (comp[0] == COMPARE_ALWAYS || comp[1] == COMPARE_ALWAYS)) return;
for (int i = 0; i < 2; i++)
{
int one = i;
int other = 1-i;
switch(op) {
case ALPHAOP_XOR:
if (comp[one] == COMPARE_ALWAYS && comp[other] == COMPARE_NEVER) return;
break;
case ALPHAOP_XNOR:
if (comp[one] == COMPARE_ALWAYS && comp[other] == COMPARE_ALWAYS) return;
if (comp[one] == COMPARE_ALWAYS && comp[other] == COMPARE_NEVER) return;
break;
}
}
//Ok, didn't get to do the easy way out :P
// do the general way
WRITE(p,"float res0, res1;\n");
WriteAlphaCompare(p, 0, bpmem.alphaFunc.comp0);
WriteAlphaCompare(p, 1, bpmem.alphaFunc.comp1);
WRITE(p,"res0 = max(res0, 0);\n");
WRITE(p,"res1 = max(res1, 0);\n");
//probably should use lookup textures for some of these :P
switch(bpmem.alphaFunc.logic) {
case ALPHAOP_AND: // if both are 0
WRITE(p,"clip(-(res0+res1)+%f);\n",epsilon);
break;
case ALPHAOP_OR: //if either is 0
WRITE(p,"clip(-res0*res1+%f);\n",epsilon*epsilon);
break;
case ALPHAOP_XOR:
//hmm, this might work:
WRITE(p,"res0=(res0>0?1:0)-.5;\n");
WRITE(p,"res1=(res1>0?1:0)-.5;\n");
WRITE(p,"clip(-res0*res1);\n",epsilon);
break;
case ALPHAOP_XNOR:
WRITE(p,"res0=(res0>0?1:0)-.5;\n");
WRITE(p,"res1=(res1>0?1:0)-.5;\n");
WRITE(p,"clip(res0*res1);\n",epsilon);
break;
}
}

30
Source/Plugins/Plugin_VideoDX9/Src/PixelShader.h
View File

@ -1,30 +0,0 @@
// 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/
#pragma once
#include "Common.h"
#include "D3DShader.h"
#define PS_CONST_COLORS 0
#define PS_CONST_KCOLORS 4
#define PS_CONST_CONSTALPHA 8
#define PS_CONST_ALPHAREF 9 // x,y
#define PS_CONST_INDMTXSTART 10
#define PS_CONST_INDSIZE 2
const char *GeneratePixelShader();

82
Source/Plugins/Plugin_VideoDX9/Src/PixelShaderCache.cpp
View File

@ -16,60 +16,44 @@
// http://code.google.com/p/dolphin-emu/
#include "D3DBase.h"
#include "D3DShader.h"
#include "Statistics.h"
#include "Utils.h"
#include "Profiler.h"
#include "PixelShaderGen.h"
#include "PixelShaderManager.h"
#include "PixelShaderCache.h"
#include "VertexLoader.h"
#include "BPMemory.h"
#include "XFMemory.h"
PShaderCache::PSCache PShaderCache::pshaders;
PixelShaderCache::PSCache PixelShaderCache::PixelShaders;
//I hope we don't get too many hash collisions :p
//all these magic numbers are primes, it should help a bit
tevhash GetCurrentTEV()
void SetPSConstant4f(int const_number, float f1, float f2, float f3, float f4)
{
u32 hash = bpmem.genMode.numindstages + bpmem.genMode.numtevstages*11 + bpmem.genMode.numtexgens*8*17;
for (int i = 0; i < (int)bpmem.genMode.numtevstages+1; i++)
{
hash = _rotl(hash,3) ^ (bpmem.combiners[i].colorC.hex*13);
hash = _rotl(hash,7) ^ ((bpmem.combiners[i].alphaC.hex&0xFFFFFFFC)*3);
hash = _rotl(hash,9) ^ xfregs.texcoords[i].texmtxinfo.projection*451;
}
for (int i = 0; i < (int)bpmem.genMode.numtevstages/2+1; i++)
{
hash = _rotl(hash,13) ^ (bpmem.tevorders[i].hex*7);
}
for (int i = 0; i < 8; i++)
{
hash = _rotl(hash,3) ^ bpmem.tevksel[i].swap1;
hash = _rotl(hash,3) ^ bpmem.tevksel[i].swap2;
}
hash ^= bpmem.dstalpha.enable ^ 0xc0debabe;
hash = _rotl(hash,4) ^ bpmem.alphaFunc.comp0*7;
hash = _rotl(hash,4) ^ bpmem.alphaFunc.comp1*13;
hash = _rotl(hash,4) ^ bpmem.alphaFunc.logic*11;
return hash;
const float f[4] = {f1, f2, f3, f4};
D3D::dev->SetPixelShaderConstantF(const_number, f, 1);
}
void SetPSConstant4fv(int const_number, const float *f)
{
D3D::dev->SetPixelShaderConstantF(const_number, f, 1);
}
void PShaderCache::Init()
void PixelShaderCache::Init()
{
}
void PShaderCache::Shutdown()
void PixelShaderCache::Shutdown()
{
PSCache::iterator iter = pshaders.begin();
for (;iter!=pshaders.end();iter++)
PSCache::iterator iter = PixelShaders.begin();
for (; iter != PixelShaders.end(); iter++)
iter->second.Destroy();
pshaders.clear();
PixelShaders.clear();
}
void PShaderCache::SetShader()
void PixelShaderCache::SetShader()
{
if (D3D::GetShaderVersion() < 2)
return; // we are screwed
@ -77,12 +61,13 @@ void PShaderCache::SetShader()
static LPDIRECT3DPIXELSHADER9 lastShader = 0;
DVSTARTPROFILE();
tevhash currentHash = GetCurrentTEV();
PIXELSHADERUID uid;
GetPixelShaderId(uid, PixelShaderManager::GetTextureMask(), false, false);
PSCache::iterator iter;
iter = pshaders.find(currentHash);
iter = PixelShaders.find(uid);
if (iter != pshaders.end())
if (iter != PixelShaders.end())
{
iter->second.frameCount = frameCount;
PSCacheEntry &entry = iter->second;
@ -94,40 +79,39 @@ void PShaderCache::SetShader()
return;
}
const char *code = GeneratePixelShader();
LPDIRECT3DPIXELSHADER9 shader = D3D::CompilePShader(code, int(strlen(code)));
const char *code = GeneratePixelShader(PixelShaderManager::GetTextureMask(), false, false);
LPDIRECT3DPIXELSHADER9 shader = D3D::CompilePixelShader(code, (int)(strlen(code)));
if (shader)
{
//Make an entry in the table
PSCacheEntry newentry;
newentry.shader = shader;
newentry.frameCount = frameCount;
pshaders[currentHash] = newentry;
PixelShaders[uid] = newentry;
INCSTAT(stats.numPixelShadersCreated);
SETSTAT(stats.numPixelShadersAlive, (int)PixelShaders.size());
}
D3D::dev->SetPixelShader(shader);
INCSTAT(stats.numPixelShadersCreated);
SETSTAT(stats.numPixelShadersAlive, (int)pshaders.size());
}
void PShaderCache::Cleanup()
void PixelShaderCache::Cleanup()
{
PSCache::iterator iter;
iter = pshaders.begin();
while (iter != pshaders.end())
iter = PixelShaders.begin();
while (iter != PixelShaders.end())
{
PSCacheEntry &entry = iter->second;
if (entry.frameCount < frameCount-30)
{
entry.Destroy();
iter = pshaders.erase(iter);
iter = PixelShaders.erase(iter);
}
else
{
iter++;
}
}
SETSTAT(stats.numPixelShadersAlive, (int)pshaders.size());
}
SETSTAT(stats.numPixelShadersAlive, (int)PixelShaders.size());
}

10
Source/Plugins/Plugin_VideoDX9/Src/PixelShaderCache.h
View File

@ -22,14 +22,14 @@
#include <map>
#include "PixelShader.h"
#include "VertexShader.h"
#include "PixelShaderGen.h"
#include "VertexShaderGen.h"
typedef u32 tevhash;
tevhash GetCurrentTEV();
class PShaderCache
class PixelShaderCache
{
struct PSCacheEntry
{
@ -48,9 +48,9 @@ class PShaderCache
}
};
typedef std::map<tevhash, PSCacheEntry> PSCache;
typedef std::map<PIXELSHADERUID, PSCacheEntry> PSCache;
static PSCache pshaders;
static PSCache PixelShaders;
public:
static void Init();

205
Source/Plugins/Plugin_VideoDX9/Src/Render.cpp
View File

@ -15,6 +15,7 @@
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#include <list>
#include <d3dx9.h>
#include "Common.h"
@ -30,16 +31,20 @@
#include "D3DPostprocess.h"
#include "D3DUtil.h"
#include "VertexShaderManager.h"
#include "VertexShaderCache.h"
#include "PixelShaderManager.h"
#include "VertexShaderCache.h"
#include "PixelShaderCache.h"
#include "TextureCache.h"
#include "Utils.h"
#include "EmuWindow.h"
#include <list>
float Renderer::m_x;
float Renderer::m_y;
float Renderer::m_width;
float Renderer::m_height;
float Renderer::xScale;
float Renderer::yScale;
float Renderer::m_x,Renderer::m_y,Renderer::m_width, Renderer::m_height, Renderer::xScale,Renderer::yScale;
std::vector<LPDIRECT3DBASETEXTURE9> Renderer::m_Textures;
DWORD Renderer::m_RenderStates[MaxRenderStates+46];
@ -50,18 +55,14 @@ DWORD Renderer::m_FVF;
#define NUMWNDRES 6
extern int g_Res[NUMWNDRES][2];
struct Message
{
Message(const std::string &msg, u32 dw) : message( msg ), dwTimeStamp( dw )
{
}
Message(const std::string &msg, u32 dw) : message(msg), dwTimeStamp(dw) { }
std::string message;
u32 dwTimeStamp;
};
static std::list<Message> s_listMsgs;
static std::list<Message> s_listMsgs;
void Renderer::Init(SVideoInitialize &_VideoInitialize)
{
@ -79,32 +80,44 @@ void Renderer::Init(SVideoInitialize &_VideoInitialize)
xScale = 640.0f / (float)vp.Width;
yScale = 480.0f / (float)vp.Height;
D3DXMATRIX mtx;
D3DXMatrixIdentity(&mtx);
D3D::dev->SetTransform(D3DTS_VIEW, &mtx);
D3D::dev->SetTransform(D3DTS_WORLD, &mtx);
float width = (float)D3D::GetDisplayWidth();
float height = (float)D3D::GetDisplayHeight();
xScale = width/640.0f;
yScale = height/480.0f;
/*
RECT rc =
{
(LONG)(m_x*xScale),
(LONG)(m_y*yScale),
(LONG)(m_width*xScale),
(LONG)(m_height*yScale)
};
*/
D3D::font.Init();
Initialize();
}
void Renderer::Shutdown(void)
void Renderer::Shutdown()
{
D3D::font.Shutdown();
D3D::EndFrame();
D3D::Close();
}
void Renderer::Initialize(void)
void Renderer::Initialize()
{
m_FVF = 0;
m_Textures.reserve(MaxTextureStages);
for (int i = 0; i < MaxTextureStages; i++)
m_Textures.push_back(NULL);
m_Textures.reserve( MaxTextureStages );
for ( int i = 0; i < MaxTextureStages; i++ )
{
m_Textures.push_back( NULL );
}
for (int i=0; i<8; i++)
{
for (int i = 0; i < 8; i++)
D3D::dev->SetSamplerState(i, D3DSAMP_MAXANISOTROPY, 16);
}
ReinitView();
Postprocess::Initialize();
Postprocess::BeginFrame();
@ -123,15 +136,15 @@ void Renderer::ProcessMessages()
int left = 25, top = 15;
std::list<Message>::iterator it = s_listMsgs.begin();
while( it != s_listMsgs.end() )
while(it != s_listMsgs.end())
{
int time_left = (int)(it->dwTimeStamp - timeGetTime());
int alpha = 255;
if(time_left<1024)
if (time_left<1024)
{
alpha=time_left>>2;
if(time_left<0) alpha=0;
if (time_left<0) alpha=0;
}
alpha <<= 24;
@ -162,28 +175,9 @@ void dumpMatrix(D3DXMATRIX &mtx)
}
}
void Renderer::ReinitView()
void Renderer::SwapBuffers()
{
D3DXMATRIX mtx;
D3DXMatrixIdentity(&mtx);
D3D::dev->SetTransform(D3DTS_VIEW,&mtx);
D3D::dev->SetTransform(D3DTS_WORLD,&mtx);
float width = (float)D3D::GetDisplayWidth();
float height = (float)D3D::GetDisplayHeight();
xScale = width/640.0f;
yScale = height/480.0f;
RECT rc =
{
(LONG)(m_x*xScale), (LONG)(m_y*yScale), (LONG)(m_width*xScale), (LONG)(m_height*yScale)
};
}
void Renderer::SwapBuffers(void)
{
// center window again
// Center window again.
if (EmuWindow::GetParentWnd())
{
RECT rcWindow;
@ -192,7 +186,7 @@ void Renderer::SwapBuffers(void)
int width = rcWindow.right - rcWindow.left;
int height = rcWindow.bottom - rcWindow.top;
::MoveWindow(EmuWindow::GetWnd(), 0,0,width,height, FALSE);
::MoveWindow(EmuWindow::GetWnd(), 0, 0, width, height, FALSE);
// nBackbufferWidth = width;
// nBackbufferHeight = height;
}
@ -230,11 +224,11 @@ void Renderer::SwapBuffers(void)
ProcessMessages();
#if defined(DVPROFILE)
if( g_bWriteProfile ) {
if (g_bWriteProfile) {
//g_bWriteProfile = 0;
static int framenum = 0;
const int UPDATE_FRAMES = 8;
if( ++framenum >= UPDATE_FRAMES ) {
if (++framenum >= UPDATE_FRAMES) {
DVProfWrite("prof.txt", UPDATE_FRAMES);
DVProfClear();
framenum = 0;
@ -248,10 +242,9 @@ void Renderer::SwapBuffers(void)
//clean out old stuff from caches
frameCount++;
PShaderCache::Cleanup();
VShaderCache::Cleanup();
PixelShaderCache::Cleanup();
VertexShaderCache::Cleanup();
TextureCache::Cleanup();
//DListCache::Cleanup();
//////////////////////////////////////////////////////////////////////////
//Begin new frame
@ -272,7 +265,8 @@ void Renderer::SwapBuffers(void)
rc.bottom = (LONG)m_height;
D3D::dev->SetScissorRect(&rc);
u32 clearColor = (bpmem.clearcolorAR<<16)|bpmem.clearcolorGB;
D3D::dev->Clear(0, 0, D3DCLEAR_TARGET, 0x101010, 0, 0);
u32 clearColor = (bpmem.clearcolorAR << 16) | bpmem.clearcolorGB;
// clearColor |= 0x003F003F;
// D3D::BeginFrame(true,clearColor,1.0f);
D3D::BeginFrame(false, clearColor, 1.0f);
@ -285,6 +279,7 @@ void Renderer::SwapBuffers(void)
D3D::font.SetRenderStates(); //compatibility with low end cards
}
/*
void Renderer::SetViewport(float* _Viewport)
{
Viewport* pViewport = (Viewport*)_Viewport;
@ -300,13 +295,13 @@ void Renderer::SetViewport(float* _Viewport)
if (y < 0.0f) y = 0.0f;
if (x > 640.0f) x = 639.0f;
if (y > 480.0f) y = 479.0f;
if (w < 0) w=1;
if (h < 0) h=1;
if (w < 0) w = 0;
if (h < 0) h = 0;
if (x+w > 640.0f) w=640-x;
if (y+h > 480.0f) h=480-y;
//x=y=0;
//if(w>0.0f) w=0.0f;
//if(h<0.0f) h=0.0f;
//if (w>0.0f) w=0.0f;
//if (h<0.0f) h=0.0f;
vp.X = (DWORD)(x*xScale);
vp.Y = (DWORD)(y*yScale);
@ -321,9 +316,12 @@ void Renderer::SetViewport(float* _Viewport)
D3D::dev->SetViewport(&vp);
}
*/
void Renderer::SetScissorBox(RECT &rc)
void Renderer::SetScissorRect()
{
RECT rc = {0,0,0,0}; // FIXX
rc.left = (int)(rc.left * xScale);
rc.top = (int)(rc.top * yScale);
rc.right = (int)(rc.right * xScale);
@ -334,6 +332,7 @@ void Renderer::SetScissorBox(RECT &rc)
g_VideoInitialize.pLog("SCISSOR ERROR", FALSE);
}
/*
void Renderer::SetProjection(float* pMatrix, int constantIndex)
{
D3DXMATRIX mtx;
@ -349,10 +348,17 @@ void Renderer::SetProjection(float* pMatrix, int constantIndex)
mtx.m[1][2] = pMatrix[3];
mtx.m[1][3] = 0; // +0.5f/m_height; <-- fix d3d pixel center?
<<<<<<< .mine
mtx.m[0][2] = 0.0f;
mtx.m[1][2] = 0.0f;
mtx.m[2][2] = -(1 - pMatrix[4]);
mtx.m[3][2] = pMatrix[5];
=======
mtx.m[2][0] = 0.0f;
mtx.m[2][1] = 0.0f;
mtx.m[2][2] = -(1.0f - pMatrix[4]);
mtx.m[2][3] = pMatrix[5]; // Problematic in OGL
>>>>>>> .r2480
mtx.m[3][0] = 0.0f;
mtx.m[3][1] = 0.0f;
@ -384,54 +390,97 @@ void Renderer::SetProjection(float* pMatrix, int constantIndex)
mtx.m[3][3] = 1.0f;
}
D3D::dev->SetVertexShaderConstantF(constantIndex, mtx, 4);
}
}*/
void Renderer::SetTexture( DWORD Stage, LPDIRECT3DBASETEXTURE9 pTexture )
void Renderer::SetTexture(DWORD Stage, LPDIRECT3DBASETEXTURE9 pTexture)
{
if ( m_Textures[Stage] != pTexture )
if (m_Textures[Stage] != pTexture)
{
m_Textures[Stage] = pTexture;
D3D::dev->SetTexture( Stage, pTexture );
D3D::dev->SetTexture(Stage, pTexture);
}
}
void Renderer::SetFVF( DWORD FVF )
void Renderer::SetFVF(DWORD FVF)
{
if ( m_FVF != FVF )
if (m_FVF != FVF)
{
m_FVF = FVF;
D3D::dev->SetFVF( FVF );
D3D::dev->SetFVF(FVF);
}
}
void Renderer::SetRenderState( D3DRENDERSTATETYPE State, DWORD Value )
void Renderer::SetRenderState(D3DRENDERSTATETYPE State, DWORD Value)
{
if ( m_RenderStates[State] != Value )
if (m_RenderStates[State] != Value)
{
m_RenderStates[State] = Value;
D3D::dev->SetRenderState( State, Value );
D3D::dev->SetRenderState(State, Value);
}
}
void Renderer::SetTextureStageState( DWORD Stage, D3DTEXTURESTAGESTATETYPE Type, DWORD Value )
void Renderer::SetTextureStageState(DWORD Stage, D3DTEXTURESTAGESTATETYPE Type, DWORD Value)
{
if ( m_TextureStageStates[Stage][Type] != Value )
if (m_TextureStageStates[Stage][Type] != Value)
{
m_TextureStageStates[Stage][Type] = Value;
D3D::dev->SetTextureStageState( Stage, Type, Value );
D3D::dev->SetTextureStageState(Stage, Type, Value);
}
}
void Renderer::SetSamplerState( DWORD Sampler, D3DSAMPLERSTATETYPE Type, DWORD Value )
void Renderer::SetSamplerState(DWORD Sampler, D3DSAMPLERSTATETYPE Type, DWORD Value)
{
if ( m_SamplerStates[Sampler][Type] != Value )
if (m_SamplerStates[Sampler][Type] != Value)
{
m_SamplerStates[Sampler][Type] = Value;
D3D::dev->SetSamplerState( Sampler, Type, Value );
D3D::dev->SetSamplerState(Sampler, Type, Value);
}
}
// Called from VertexShaderManager
void UpdateViewport()
{
// reversed gxsetviewport(xorig, yorig, width, height, nearz, farz)
// [0] = width/2
// [1] = height/2
// [2] = 16777215 *(farz - nearz)
// [3] = xorig + width/2 + 342
// [4] = yorig + height/2 + 342
// [5] = 16777215 * farz
/*INFO_LOG("view: topleft=(%f,%f), wh=(%f,%f), z=(%f,%f)\n",
rawViewport[3]-rawViewport[0]-342, rawViewport[4]+rawViewport[1]-342,
2 * rawViewport[0], 2 * rawViewport[1],
(rawViewport[5] - rawViewport[2]) / 16777215.0f, rawViewport[5] / 16777215.0f);*/
D3DVIEWPORT9 vp;
// Keep aspect ratio at 4:3
// rawViewport[0] = 320, rawViewport[1] = -240
int scissorXOff = bpmem.scissorOffset.x * 2 - 342;
int scissorYOff = bpmem.scissorOffset.y * 2 - 342;
float fourThree = 4.0f / 3.0f;
float wAdj, hAdj;
float actualRatiow, actualRatioh;
int overfl;
int xoffs = 0;
int yoffs = 0;
int wid, hei, actualWid, actualHei;
int winw = 640;
int winh = 480;
float ratio = (float)winw / (float)winh / fourThree;
vp.MinZ = (xfregs.rawViewport[5] - xfregs.rawViewport[2])/16777215.0f;
vp.MaxZ = xfregs.rawViewport[5]/16777215.0f;
wid = ceil(fabs(2 * xfregs.rawViewport[0]));
hei = ceil(fabs(2 * xfregs.rawViewport[1]));
vp.X = (int)(xfregs.rawViewport[3] - xfregs.rawViewport[0] - 342 - scissorXOff) + xoffs;
vp.Y = (int)(xfregs.rawViewport[4] - xfregs.rawViewport[1] - 342 - scissorYOff) + yoffs;
vp.Width = wid;
vp.Height = hei;
D3D::dev->SetViewport(&vp);
}

26
Source/Plugins/Plugin_VideoDX9/Src/Render.h
View File

@ -48,19 +48,19 @@ public:
static void Init(SVideoInitialize &_VideoInitialize);
static void Shutdown();
static void Initialize(void);
static void Initialize();
// must be called if the window size has changed
static void ReinitView(void);
static void ReinitView();
static void SwapBuffers(void);
static void SwapBuffers();
static float GetXScale() {return xScale;}
static float GetYScale() {return yScale;}
static float GetXScale() { return xScale; }
static float GetYScale() { return yScale; }
static void SetScissorBox(RECT &rc);
static void SetViewport(float* _Viewport);
static void SetProjection(float* _pProjection, int constantIndex = -1);
static void SetScissorRect();
// static void SetViewport(float* _Viewport);
// static void SetProjection(float* _pProjection, int constantIndex = -1);
// The little status display.
static void AddMessage(const std::string &message, unsigned int ms);
@ -68,11 +68,11 @@ public:
static void RenderText(const std::string &text, int left, int top, unsigned int color);
// The following are "filtered" versions of the corresponding D3Ddev-> functions.
static void SetTexture( DWORD Stage, IDirect3DBaseTexture9 *pTexture );
static void SetFVF( DWORD FVF );
static void SetRenderState( D3DRENDERSTATETYPE State, DWORD Value );
static void SetTextureStageState( DWORD Stage, D3DTEXTURESTAGESTATETYPE Type,DWORD Value );
static void SetSamplerState( DWORD Sampler, D3DSAMPLERSTATETYPE Type, DWORD Value );
static void SetTexture(DWORD Stage, IDirect3DBaseTexture9 *pTexture);
static void SetFVF(DWORD FVF);
static void SetRenderState(D3DRENDERSTATETYPE State, DWORD Value);
static void SetTextureStageState(DWORD Stage, D3DTEXTURESTAGESTATETYPE Type, DWORD Value);
static void SetSamplerState(DWORD Sampler, D3DSAMPLERSTATETYPE Type, DWORD Value);
};
#endif // __H_RENDER__

372
Source/Plugins/Plugin_VideoDX9/Src/TransformEngine.cpp
View File

@ -1,372 +0,0 @@
// 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 <fvec.h>
#include "Common.h"
#include "Profiler.h"
// #include "Globals.h"
#include "Vec3.h"
#include "TransformEngine.h"
#include "VertexManager.h"
#include "VertexLoader.h"
#include "BPStructs.h"
#include "XFStructs.h"
#include "Utils.h"
#include "RGBAFloat.h"
float *CTransformEngine::m_pPosMatrix;
float *CTransformEngine::m_pNormalMatrix;
float *CTransformEngine::m_pTexMatrix[8];
float *CTransformEngine::m_pTexPostMatrix[8];
const Light *GetLight(int i)
{
return (const Light *)(xfmem + XFMEM_LIGHTS) + i;
}
float DoLighting(const Light *light, const LitChannel &chan, const Vec3 &pos, const Vec3 &normal)
{
float val;
if (chan.attnfunc == 0 || chan.attnfunc == 2) //no attn
{
Vec3 ldir = (Vec3(light->dpos) - pos);
val = ldir.normalized() * normal;
}
else
{
float aattn = 0;
float d;
float mul = 1.0f;
if (chan.attnfunc == 3)
{
Vec3 ldir = (Vec3(light->dpos) - pos);
d = ldir.length();
Vec3 ldirNorm = ldir / d; //normalize
float l = ldirNorm * normal;
aattn = Vec3(light->ddir) * ldirNorm;
mul = l;
}
else if (chan.attnfunc == 1)
{
d = aattn = Vec3(light->shalfangle) * normal;
mul = (Vec3(light->sdir) * normal > 0) ? (normal * Vec3(light->shalfangle)) : 0;
if (mul < 0)
mul = 0;
}
float spot = (light->a2*aattn*aattn + light->a1*aattn + light->a0);
float dist = 1.0f/(light->k2*d*d + light->k1*d + light->k0);
if (spot<0)
spot=0;
val = mul * spot * dist;
}
if (val < 0 && chan.diffusefunc == 2) // clamp
val = 0;
return val;
}
void VtxMulMtx43T(Vec3 &out, const Vec3 &in, const float pMatrix[12])
{
out.x = in.x * pMatrix[0] + in.y * pMatrix[1] + in.z * pMatrix[2] + 1 * pMatrix[3];
out.y = in.x * pMatrix[4] + in.y * pMatrix[5] + in.z * pMatrix[6] + 1 * pMatrix[7];
out.z = in.x * pMatrix[8] + in.y * pMatrix[9] + in.z * pMatrix[10] + 1 * pMatrix[11];
}
void VtxMulMtx43(Vec3 &out, const Vec3 &in, const float pMatrix[12])
{
VtxMulMtx43T(out,in,pMatrix);
//TODO(XK): Turns out that SSE2 computations are slower... Can anyone do
// anything about it?
/*
F32vec4 a(in.x, in.y, in.z, 1), b(pMatrix[0], pMatrix[1], pMatrix[2], pMatrix[3]);
out.x = add_horizontal(a * b);
b[0] = pMatrix[4]; b[1] = pMatrix[5]; b[2] = pMatrix[6]; b[3] = pMatrix[7];
out.y = add_horizontal(a * b);
b[0] = pMatrix[8]; b[1] = pMatrix[9]; b[2] = pMatrix[10]; b[3] = pMatrix[11];
out.z = add_horizontal(a * b);
*/
}
void VtxMulMtx42(Vec3 &out, const Vec3 &in, const float pMatrix[8])
{
out.x = in.x * pMatrix[0] + in.y * pMatrix[1] + in.z * pMatrix[2] + 1 * pMatrix[3];
out.y = in.x * pMatrix[4] + in.y * pMatrix[5] + in.z * pMatrix[6] + 1 * pMatrix[7];
}
void VtxMulMtx33(Vec3 &out, const Vec3 &in, const float pMatrix[9])
{
out.x = in.x * pMatrix[0] + in.y * pMatrix[1] + in.z * pMatrix[2];
out.y = in.x * pMatrix[3] + in.y * pMatrix[4] + in.z * pMatrix[5];
out.z = in.x * pMatrix[6] + in.y * pMatrix[7] + in.z * pMatrix[8];
}
void CTransformEngine::TransformVertices(int _numVertices, const DecodedVArray *varray, D3DVertex *vbuffer)
{
if (vbuffer == 0)
{
MessageBox(0,"TransformVertices : vbuffer == 0","WTF",0);
}
DVSTARTPROFILE();
RGBAFloat lightColors[8];
RGBAFloat lightVals[8];
RGBAFloat chans[2];
u32 components = varray->GetComponents();
// TODO: only for active lights
for (int i=0; i<8; i++)
lightColors[i].convert_GC(GetLight(i)->color);
for (int i=0; i<_numVertices; i++)
{
//////////////////////////////////////////////////////////////////////////
//Step 1: xform position and normal
//////////////////////////////////////////////////////////////////////////
Vec3 OrigPos = varray->GetPos(i);
if (varray->hasPosMatIdx)
{
int index = varray->GetPosMtxInd(i);
SetPosNormalMatrix(
(float*)xfmem + (index & 63) * 4, //CHECK
(float*)xfmem + 0x400 + 3 * (index & 31)); //CHECK
}
for (int j = 0; j < 8; j++)
{
if (varray->hasTexMatIdx[j])
{
float *flipmem = (float *)xfmem;
int index = varray->GetTexMtxInd(j, i);
SetTexMatrix(j, flipmem + index * 4);
}
}
Vec3 TempPos;
// m_pPosMatrix can be switched out, through matrixindex vertex components
VtxMulMtx43(TempPos, OrigPos, m_pPosMatrix);
Vec3 TempNormal;
Vec3 OrigNormal;
if (varray->hasNrm)
{
OrigNormal = varray->GetNormal(0, i);
VtxMulMtx33(TempNormal, OrigNormal, m_pNormalMatrix);
TempNormal.normalize();
}
else
{
OrigNormal.setZero();
TempNormal.setZero();
}
//////////////////////////////////////////////////////////////////////////
//Step 2: Light!
//////////////////////////////////////////////////////////////////////////
//find all used lights
u32 lightMask =
xfregs.colChans[0].color.GetFullLightMask() | xfregs.colChans[0].alpha.GetFullLightMask() |
xfregs.colChans[1].color.GetFullLightMask() | xfregs.colChans[1].alpha.GetFullLightMask();
float r0=0,g0=0,b0=0,a0=0;
//go through them and compute the lit colors
//Sum lighting for both two color channels if they're active
for (int j = 0; j < (int)bpmem.genMode.numcolchans; j++)
{
RGBAFloat material;
RGBAFloat lightSum(0,0,0,0);
bool hasColorJ = (components & (VertexLoader::VB_HAS_COL0 << j)) != 0;
//get basic material color from appropriate sources (this would compile nicely!:)
if (xfregs.colChans[j].color.matsource == GX_SRC_REG)
material.convertRGB_GC(xfregs.colChans[j].matColor);
else
{
if (hasColorJ)
material.convertRGB(varray->GetColor(j, i));
else
material.r=material.g=material.b=1.0f;
}
if (xfregs.colChans[j].alpha.matsource == GX_SRC_REG)
material.convertA_GC(xfregs.colChans[j].matColor);
else
{
if (hasColorJ)
material.convertA(varray->GetColor(j, i));
else
material.a=1.0f;
}
//combine together the light values from the lights that affect the color
if (xfregs.colChans[j].color.enablelighting)
{
//choose ambient source and start our lightsum accumulator with its value..
if (xfregs.colChans[j].color.ambsource == GX_SRC_REG)
lightSum.convertRGB_GC(xfregs.colChans[j].ambColor); //ambient
else
{
if (hasColorJ)
lightSum.convertRGB(varray->GetColor(j, i));
else
{
lightSum.r=0.0f;lightSum.g=0.0f;lightSum.b=0.0f;
}
}
//accumulate light colors
int cmask = xfregs.colChans[j].color.GetFullLightMask();
for (int l=0; l<8; l++)
{
if (cmask&1)
{
float val = DoLighting(GetLight(l), xfregs.colChans[j].color, TempPos, TempNormal);
float r = lightColors[l].r * val;
float g = lightColors[l].g * val;
float b = lightColors[l].b * val;
lightSum.r += r;
lightSum.g += g;
lightSum.b += b;
}
cmask >>= 1;
}
}
else
{
lightSum.r = lightSum.g = lightSum.b = 1.0f;
}
//combine together the light values from the lights that affect alpha (should be rare)
if (xfregs.colChans[j].alpha.enablelighting)
{
//choose ambient source..
if (xfregs.colChans[j].alpha.ambsource==GX_SRC_REG)
lightSum.convertA_GC(xfregs.colChans[j].ambColor);
else
{
if (hasColorJ)
lightSum.convertA(varray->GetColor(j, i));
else
lightSum.a=0.0f;
}
//accumulate light alphas
int amask = xfregs.colChans[j].alpha.GetFullLightMask();
for (int l = 0; l < 8; l++)
{
if (amask&1)
{
float val = DoLighting(GetLight(l), xfregs.colChans[j].alpha, TempPos, TempNormal);
float a = lightColors[l].a * val;
lightSum.a += a;
}
amask >>= 1;
}
}
else
{
lightSum.a=1.0f;
}
chans[j] = lightSum * material;
chans[j].clamp();
}
//////////////////////////////////////////////////////////////////////////
//Step 3: Generate texture coordinates!
//////////////////////////////////////////////////////////////////////////
Vec3 TempUVs[8];
for (int j = 0; j < xfregs.numTexGens; j++)
{
Vec3 t;
switch (xfregs.texcoords[j].texmtxinfo.sourcerow) {
case XF_SRCGEOM_INROW: t = OrigPos; break; //HACK WTFF???
case XF_SRCNORMAL_INROW: t = OrigNormal; break;
case XF_SRCCOLORS_INROW: break; //set uvs to something?
case XF_SRCBINORMAL_T_INROW: t=Vec3(0,0,0);break;
case XF_SRCBINORMAL_B_INROW: t=Vec3(0,0,0);break;
default:
{
int c = xfregs.texcoords[j].texmtxinfo.sourcerow - XF_SRCTEX0_INROW;
bool hasTCC = (components & (VertexLoader::VB_HAS_UV0 << c)) != 0;
if (c >= 0 && c <= 7 && hasTCC)
{
const DecUV &uv = varray->GetUV(c, i);
t = Vec3(uv.u, uv.v, 1);
}
}
}
Vec3 out,out2;
switch (xfregs.texcoords[j].texmtxinfo.texgentype)
{
case XF_TEXGEN_COLOR_STRGBC0:
out = Vec3(chans[0].r*255, chans[0].g*255, 1)/255.0f;
break;
case XF_TEXGEN_COLOR_STRGBC1:
out = Vec3(chans[1].r*255, chans[1].g*255, 1)/255.0f; //FIX: take color1 instead
break;
case XF_TEXGEN_REGULAR:
if (xfregs.texcoords[j].texmtxinfo.projection)
VtxMulMtx43(out, t, m_pTexMatrix[j]);
else
VtxMulMtx42(out, t, m_pTexMatrix[j]);
break;
}
if (xfregs.texcoords[j].postmtxinfo.normalize)
out.normalize();
int postMatrix = xfregs.texcoords[j].postmtxinfo.index;
float *pmtx = ((float*)xfmem) + 0x500 + postMatrix * 4; //CHECK
//multiply with postmatrix
VtxMulMtx43(TempUVs[j], out, pmtx);
}
//////////////////////////////////////////////////////////////////////////
//Step 4: Output the vertex!
//////////////////////////////////////////////////////////////////////////
for (int j = 0; j < 2; j++)
chans[j].convertToD3DColor(vbuffer[i].colors[j]);
vbuffer[i].pos = TempPos;
vbuffer[i].normal = TempNormal;
for (int j = 0; j < (int)bpmem.genMode.numtexgens; j++)
{
vbuffer[i].uv[j].u = TempUVs[j].x;
vbuffer[i].uv[j].v = TempUVs[j].y;
vbuffer[i].uv[j].w = TempUVs[j].z;
}
}
}

55
Source/Plugins/Plugin_VideoDX9/Src/TransformEngine.h
View File

@ -1,55 +0,0 @@
// 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/
#ifndef _TRANSFORMENGINE_H
#define _TRANSFORMENGINE_H
//T&L Engine
//as much work as possible will be delegated to vertex shaders later
//to take full advantage of current PC HW
#include "VertexManager.h"
#include "DecodedVArray.h"
class CTransformEngine
{
static float* m_pPosMatrix;
static float* m_pNormalMatrix;
static float* m_pTexMatrix[8];
static float* m_pTexPostMatrix[8];
public:
static void TransformVertices(int _numVertices,
const DecodedVArray *varray,
D3DVertex *vbuffer);
static void SetPosNormalMatrix(float *p, float *n)
{
m_pPosMatrix = p;
m_pNormalMatrix = n;
}
static void SetTexMatrix(int i, float *f)
{
m_pTexMatrix[i] = f;
}
static void SetTexPostMatrix(int i, float *f)
{
m_pTexPostMatrix[i] = f;
}
};
#endif

363
Source/Plugins/Plugin_VideoDX9/Src/VertexLoader.cpp
View File

@ -1,363 +0,0 @@
// 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 <stdio.h>
#include "x64Emitter.h"
#include "Common.h"
#include "LookUpTables.h"
#include "Profiler.h"
#include "VertexManager.h"
#include "VertexLoader.h"
#include "XFStructs.h"
#include "BPStructs.h"
#include "DataReader.h"
#include "DecodedVArray.h"
//these don't need to be saved
float posScale;
float tcScale[8];
int tcElements[8];
int tcFormat[8];
int colElements[2];
float tcScaleU[8];
float tcScaleV[8];
int tcIndex;
int colIndex;
u32 addr;
DecodedVArray *varray;
int ComputeVertexSize(u32 comp)
{
int size = 0;
if (comp & VertexLoader::VB_HAS_POSMTXIDX)
size += 4;
if (comp & (VertexLoader::VB_HAS_TEXMTXIDX0 | VertexLoader::VB_HAS_TEXMTXIDX1 | VertexLoader::VB_HAS_TEXMTXIDX2 | VertexLoader::VB_HAS_TEXMTXIDX3))
size += 4;
if (comp & (VertexLoader::VB_HAS_TEXMTXIDX4 | VertexLoader::VB_HAS_TEXMTXIDX5 | VertexLoader::VB_HAS_TEXMTXIDX6 | VertexLoader::VB_HAS_TEXMTXIDX7))
size += 4;
if (comp & VertexLoader::VB_HAS_NRM0)
size += 4;
if (comp & (VertexLoader::VB_HAS_NRM1 | VertexLoader::VB_HAS_NRM2)) //combine into single check for speed
size += 8;
if (comp & VertexLoader::VB_HAS_COL0)
size += 4;
if (comp & VertexLoader::VB_HAS_COL1)
size += 4;
for (int i = 0; i < 8; i++)
if (comp & (VertexLoader::VB_HAS_UV0 << i))
size += 8;
return size;
}
void VertexLoader::SetVArray(DecodedVArray *_varray)
{
varray = _varray;
}
#include "VertexLoader_MtxIndex.h"
#include "VertexLoader_Position.h"
#include "VertexLoader_Normal.h"
#include "VertexLoader_Color.h"
#include "VertexLoader_TextCoord.h"
VertexLoader g_VertexLoaders[8];
TVtxDesc VertexLoader::m_VtxDesc;
bool VertexLoader::m_DescDirty = true;
VertexLoader::VertexLoader()
{
m_numPipelineStates = 0;
m_VertexSize = 0;
m_AttrDirty = true;
VertexLoader_Normal::Init();
}
VertexLoader::~VertexLoader()
{
}
void VertexLoader::Setup()
{
if (!m_AttrDirty && !m_DescDirty)
return;
DVSTARTPROFILE();
// Reset pipeline
m_VertexSize = 0;
m_numPipelineStates = 0;
m_components = 0;
// Position Matrix Index
if (m_VtxDesc.PosMatIdx)
{
m_PipelineStates[m_numPipelineStates++] = PosMtx_ReadDirect_UByte;
m_VertexSize += 1;
m_components |= VB_HAS_POSMTXIDX;
}
// Texture matrix indices
if (m_VtxDesc.Tex0MatIdx) {m_components|=VB_HAS_TEXMTXIDX0; WriteCall(TexMtx_ReadDirect_UByte); m_VertexSize+=1;}
if (m_VtxDesc.Tex1MatIdx) {m_components|=VB_HAS_TEXMTXIDX1; WriteCall(TexMtx_ReadDirect_UByte); m_VertexSize+=1;}
if (m_VtxDesc.Tex2MatIdx) {m_components|=VB_HAS_TEXMTXIDX2; WriteCall(TexMtx_ReadDirect_UByte); m_VertexSize+=1;}
if (m_VtxDesc.Tex3MatIdx) {m_components|=VB_HAS_TEXMTXIDX3; WriteCall(TexMtx_ReadDirect_UByte); m_VertexSize+=1;}
if (m_VtxDesc.Tex4MatIdx) {m_components|=VB_HAS_TEXMTXIDX4; WriteCall(TexMtx_ReadDirect_UByte); m_VertexSize+=1;}
if (m_VtxDesc.Tex5MatIdx) {m_components|=VB_HAS_TEXMTXIDX5; WriteCall(TexMtx_ReadDirect_UByte); m_VertexSize+=1;}
if (m_VtxDesc.Tex6MatIdx) {m_components|=VB_HAS_TEXMTXIDX6; WriteCall(TexMtx_ReadDirect_UByte); m_VertexSize+=1;}
if (m_VtxDesc.Tex7MatIdx) {m_components|=VB_HAS_TEXMTXIDX7; WriteCall(TexMtx_ReadDirect_UByte); m_VertexSize+=1;}
// Position
switch (m_VtxDesc.Position)
{
case NOT_PRESENT: {_assert_msg_(0,"Vertex descriptor without position!","WTF?");} break;
case DIRECT:
{
int SizePro = 0;
switch (m_VtxAttr.PosFormat)
{
case FORMAT_UBYTE: SizePro=1; WriteCall(Pos_ReadDirect_UByte); break;
case FORMAT_BYTE: SizePro=1; WriteCall(Pos_ReadDirect_Byte); break;
case FORMAT_USHORT: SizePro=2; WriteCall(Pos_ReadDirect_UShort); break;
case FORMAT_SHORT: SizePro=2; WriteCall(Pos_ReadDirect_Short); break;
case FORMAT_FLOAT: SizePro=4; WriteCall(Pos_ReadDirect_Float); break;
default: _assert_(0); break;
}
if (m_VtxAttr.PosElements == 1)
m_VertexSize += SizePro * 3;
else
m_VertexSize += SizePro * 2;
}
break;
case INDEX8:
m_VertexSize+=1;
switch (m_VtxAttr.PosFormat)
{
case FORMAT_UBYTE: WriteCall(Pos_ReadIndex8_UByte); break; //WTF?
case FORMAT_BYTE: WriteCall(Pos_ReadIndex8_Byte); break;
case FORMAT_USHORT: WriteCall(Pos_ReadIndex8_UShort); break;
case FORMAT_SHORT: WriteCall(Pos_ReadIndex8_Short); break;
case FORMAT_FLOAT: WriteCall(Pos_ReadIndex8_Float); break;
default: _assert_(0); break;
}
break;
case INDEX16:
m_VertexSize+=2;
switch (m_VtxAttr.PosFormat)
{
case FORMAT_UBYTE: WriteCall(Pos_ReadIndex16_UByte); break;
case FORMAT_BYTE: WriteCall(Pos_ReadIndex16_Byte); break;
case FORMAT_USHORT: WriteCall(Pos_ReadIndex16_UShort); break;
case FORMAT_SHORT: WriteCall(Pos_ReadIndex16_Short); break;
case FORMAT_FLOAT: WriteCall(Pos_ReadIndex16_Float); break;
default: _assert_(0); break;
}
break;
}
// Normals
if (m_VtxDesc.Normal != NOT_PRESENT)
{
VertexLoader_Normal::index3 = m_VtxAttr.NormalIndex3 ? true : false;
unsigned int uSize = VertexLoader_Normal::GetSize(m_VtxDesc.Normal, m_VtxAttr.NormalFormat, m_VtxAttr.NormalElements);
TPipelineFunction pFunc = VertexLoader_Normal::GetFunction(m_VtxDesc.Normal, m_VtxAttr.NormalFormat, m_VtxAttr.NormalElements);
if (pFunc == 0)
{
char temp[256];
sprintf(temp,"%i %i %i", m_VtxDesc.Normal, m_VtxAttr.NormalFormat, m_VtxAttr.NormalElements);
MessageBox(0,"VertexLoader_Normal::GetFunction returned zero!",temp,0);
}
WriteCall(pFunc);
m_VertexSize += uSize;
int m_numNormals = (m_VtxAttr.NormalElements == 1) ? NRM_THREE : NRM_ONE;
m_components |= VB_HAS_NRM0;
if (m_numNormals == NRM_THREE)
m_components |= VB_HAS_NRM1 | VB_HAS_NRM2;
}
// Colors
int col[2] = {m_VtxDesc.Color0, m_VtxDesc.Color1};
for (int i = 0; i < 2; i++)
SetupColor(i,col[i], m_VtxAttr.color[i].Comp, m_VtxAttr.color[i].Elements);
// TextureCoord
// Since m_VtxDesc.Text7Coord is broken across a 32 bit word boundary, retrieve its value manually.
// If we didn't do this, the vertex format would be read as one bit offset from where it should be, making
// 01 become 00, and 10/11 become 01
int tc[8] = {
m_VtxDesc.Tex0Coord, m_VtxDesc.Tex1Coord, m_VtxDesc.Tex2Coord, m_VtxDesc.Tex3Coord,
m_VtxDesc.Tex4Coord, m_VtxDesc.Tex5Coord, m_VtxDesc.Tex6Coord, (m_VtxDesc.Hex >> 31) & 3
};
for (int i = 0; i < 8; i++)
SetupTexCoord(i, tc[i],
m_VtxAttr.texCoord[i].Format,
m_VtxAttr.texCoord[i].Elements,
m_VtxAttr.texCoord[i].Frac);
}
void VertexLoader::SetupColor(int num, int mode, int format, int elements)
{
m_components |= VB_HAS_COL0 << num;
switch (mode)
{
case NOT_PRESENT:
m_components &= ~(VB_HAS_COL0 << num);
break;
case DIRECT:
switch (format)
{
case FORMAT_16B_565: m_VertexSize+=2; WriteCall(Color_ReadDirect_16b_565); break;
case FORMAT_24B_888: m_VertexSize+=3; WriteCall(Color_ReadDirect_24b_888); break;
case FORMAT_32B_888x: m_VertexSize+=4; WriteCall(Color_ReadDirect_32b_888x); break;
case FORMAT_16B_4444: m_VertexSize+=2; WriteCall(Color_ReadDirect_16b_4444); break;
case FORMAT_24B_6666: m_VertexSize+=3; WriteCall(Color_ReadDirect_24b_6666); break;
case FORMAT_32B_8888: m_VertexSize+=4; WriteCall(Color_ReadDirect_32b_8888); break;
default: _assert_(0); break;
}
break;
case INDEX8:
switch (format)
{
case FORMAT_16B_565: WriteCall(Color_ReadIndex8_16b_565); break;
case FORMAT_24B_888: WriteCall(Color_ReadIndex8_24b_888); break;
case FORMAT_32B_888x: WriteCall(Color_ReadIndex8_32b_888x); break;
case FORMAT_16B_4444: WriteCall(Color_ReadIndex8_16b_4444); break;
case FORMAT_24B_6666: WriteCall(Color_ReadIndex8_24b_6666); break;
case FORMAT_32B_8888: WriteCall(Color_ReadIndex8_32b_8888); break;
default: _assert_(0); break;
}
m_VertexSize+=1;
break;
case INDEX16:
switch (format)
{
case FORMAT_16B_565: WriteCall(Color_ReadIndex16_16b_565); break;
case FORMAT_24B_888: WriteCall(Color_ReadIndex16_24b_888); break;
case FORMAT_32B_888x: WriteCall(Color_ReadIndex16_32b_888x); break;
case FORMAT_16B_4444: WriteCall(Color_ReadIndex16_16b_4444); break;
case FORMAT_24B_6666: WriteCall(Color_ReadIndex16_24b_6666); break;
case FORMAT_32B_8888: WriteCall(Color_ReadIndex16_32b_8888); break;
default: _assert_(0); break;
}
m_VertexSize+=2;
break;
}
}
void VertexLoader::SetupTexCoord(int num, int mode, int format, int elements, int _iFrac)
{
m_components |= VB_HAS_UV0 << num;
switch (mode)
{
case NOT_PRESENT:
m_components &= ~(VB_HAS_UV0 << num);
break;
case DIRECT:
{
int sizePro=0;
switch (format)
{
case FORMAT_UBYTE: sizePro = 1; WriteCall(TexCoord_ReadDirect_UByte); break;
case FORMAT_BYTE: sizePro = 1; WriteCall(TexCoord_ReadDirect_Byte); break;
case FORMAT_USHORT: sizePro = 2; WriteCall(TexCoord_ReadDirect_UShort); break;
case FORMAT_SHORT: sizePro = 2; WriteCall(TexCoord_ReadDirect_Short); break;
case FORMAT_FLOAT: sizePro = 4; WriteCall(TexCoord_ReadDirect_Float); break;
default: _assert_(0); break;
}
m_VertexSize += sizePro * (elements ? 2 : 1);
}
break;
case INDEX8:
switch (format)
{
case FORMAT_UBYTE: WriteCall(TexCoord_ReadIndex8_UByte); break;
case FORMAT_BYTE: WriteCall(TexCoord_ReadIndex8_Byte); break;
case FORMAT_USHORT: WriteCall(TexCoord_ReadIndex8_UShort); break;
case FORMAT_SHORT: WriteCall(TexCoord_ReadIndex8_Short); break;
case FORMAT_FLOAT: WriteCall(TexCoord_ReadIndex8_Float); break;
default: _assert_(0); break;
}
m_VertexSize+=1;
break;
case INDEX16:
switch (format)
{
case FORMAT_UBYTE: WriteCall(TexCoord_ReadIndex16_UByte); break;
case FORMAT_BYTE: WriteCall(TexCoord_ReadIndex16_Byte); break;
case FORMAT_USHORT: WriteCall(TexCoord_ReadIndex16_UShort); break;
case FORMAT_SHORT: WriteCall(TexCoord_ReadIndex16_Short); break;
case FORMAT_FLOAT: WriteCall(TexCoord_ReadIndex16_Float); break;
default: _assert_(0);
}
m_VertexSize+=2;
break;
}
}
void VertexLoader::WriteCall(TPipelineFunction func)
{
m_PipelineStates[m_numPipelineStates++] = func;
}
using namespace Gen;
void VertexLoader::PrepareRun()
{
posScale = shiftLookup[m_VtxAttr.PosFrac];
for (int i = 0; i < 8; i++)
{
tcScaleU[i] = shiftLookup[m_VtxAttr.texCoord[i].Frac];
tcScaleV[i] = shiftLookup[m_VtxAttr.texCoord[i].Frac];
tcElements[i] = m_VtxAttr.texCoord[i].Elements;
tcFormat[i] = m_VtxAttr.texCoord[i].Format;
}
for (int i = 0; i < 2; i++)
colElements[i] = m_VtxAttr.color[i].Elements;
varray->SetComponents(m_components);
}
void VertexLoader::RunVertices(int count)
{
DVSTARTPROFILE();
for (int v = 0; v < count; v++)
{
tcIndex = 0;
colIndex = 0;
s_texmtxread = 0;
for (int i = 0; i < m_numPipelineStates; i++)
{
m_PipelineStates[i](&m_VtxAttr);
}
varray->Next();
}
/*
This is not the bottleneck ATM, so compiling etc doesn't really help.
At least not when all we do is compile it to a list of function calls.
Should help more when we inline, but this requires the new vertex format.
Maybe later, and with smarter caching.
if (count)
{
this->m_counter = count;
((void (*)())((void*)&m_compiledCode[0]))();
}*/
}

194
Source/Plugins/Plugin_VideoDX9/Src/VertexLoader.h
View File

@ -1,194 +0,0 @@
// 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/
#ifndef _VERTEXLOADER_H
#define _VERTEXLOADER_H
int ComputeVertexSize(u32 components);
#include "CPStructs.h"
#include "DecodedVArray.h"
#include "DataReader.h"
#define LOADERDECL __cdecl
typedef void (LOADERDECL *TPipelineFunction)(const void*);
class VertexLoader
{
public:
enum
{
NRM_ZERO = 0,
NRM_ONE = 1,
NRM_THREE = 3
};
enum {
VB_HAS_POSMTXIDX =(1<<1),
VB_HAS_TEXMTXIDX0=(1<<2),
VB_HAS_TEXMTXIDX1=(1<<3),
VB_HAS_TEXMTXIDX2=(1<<4),
VB_HAS_TEXMTXIDX3=(1<<5),
VB_HAS_TEXMTXIDX4=(1<<6),
VB_HAS_TEXMTXIDX5=(1<<7),
VB_HAS_TEXMTXIDX6=(1<<8),
VB_HAS_TEXMTXIDX7=(1<<9),
VB_HAS_TEXMTXIDXALL=(0xff<<2),
//VB_HAS_POS=0, // Implied, it always has pos! don't bother testing
VB_HAS_NRM0=(1<<10),
VB_HAS_NRM1=(1<<11),
VB_HAS_NRM2=(1<<12),
VB_HAS_NRMALL=(7<<10),
VB_HAS_COL0=(1<<13),
VB_HAS_COL1=(1<<14),
VB_HAS_UV0=(1<<15),
VB_HAS_UV1=(1<<16),
VB_HAS_UV2=(1<<17),
VB_HAS_UV3=(1<<18),
VB_HAS_UV4=(1<<19),
VB_HAS_UV5=(1<<20),
VB_HAS_UV6=(1<<21),
VB_HAS_UV7=(1<<22),
VB_HAS_UVALL=(0xff<<15),
VB_HAS_UVTEXMTXSHIFT=13,
};
private:
TPipelineFunction m_PipelineStates[32];
int m_numPipelineStates;
int m_VertexSize;
int m_counter;
u32 m_components;
UVAT_group0 m_group0;
UVAT_group1 m_group1;
UVAT_group2 m_group2;
bool m_AttrDirty;
TVtxAttr m_VtxAttr; //Decoded into easy format
//common for all loaders
static TVtxDesc m_VtxDesc;
static bool m_DescDirty;
// seup the pipeline with this vertex fmt
void SetupColor(int num, int _iMode, int _iFormat, int _iElements);
void SetupTexCoord(int num, int _iMode, int _iFormat, int _iElements, int _iFrac);
public:
// constructor
VertexLoader();
~VertexLoader();
// run the pipeline
static void SetVArray(DecodedVArray *_varray);
void Setup();
void PrepareRun();
void RunVertices(int count);
void WriteCall(TPipelineFunction func);
int GetVertexSize(){return m_VertexSize;}
//VtxDesc - global
static void SetVtxDesc_Lo(u32 _iValue)
{
u64 old = m_VtxDesc.Hex;
m_VtxDesc.Hex &= ~0x1FFFF; // keep the Upper bits
m_VtxDesc.Hex |= _iValue;
if (m_VtxDesc.Hex != old)
m_DescDirty = true;
};
static void SetVtxDesc_Hi(u32 _iValue)
{
u64 old = m_VtxDesc.Hex;
m_VtxDesc.Hex &= 0x1FFFF; // keep the lower 17Bits
m_VtxDesc.Hex |= (u64)_iValue << 17;
if (m_VtxDesc.Hex != old)
m_DescDirty = true;
};
static TVtxDesc &GetVtxDesc() {return m_VtxDesc;}
void SetVAT_group0(u32 _group0)
{
if (m_group0.Hex == _group0)
return;
m_AttrDirty = true;
m_group0.Hex = _group0;
m_VtxAttr.PosElements = m_group0.PosElements;
m_VtxAttr.PosFormat = m_group0.PosFormat;
m_VtxAttr.PosFrac = m_group0.PosFrac;
m_VtxAttr.NormalElements = m_group0.NormalElements;
m_VtxAttr.NormalFormat = m_group0.NormalFormat;
m_VtxAttr.color[0].Elements = m_group0.Color0Elements;
m_VtxAttr.color[0].Comp = m_group0.Color0Comp;
m_VtxAttr.color[1].Elements = m_group0.Color1Elements;
m_VtxAttr.color[1].Comp = m_group0.Color1Comp;
m_VtxAttr.texCoord[0].Elements = m_group0.Tex0CoordElements;
m_VtxAttr.texCoord[0].Format = m_group0.Tex0CoordFormat;
m_VtxAttr.texCoord[0].Frac = m_group0.Tex0Frac;
m_VtxAttr.ByteDequant = m_group0.ByteDequant;
m_VtxAttr.NormalIndex3 = m_group0.NormalIndex3;
};
void SetVAT_group1(u32 _group1)
{
if (m_group1.Hex == _group1)
return;
m_AttrDirty = true;
m_group1.Hex = _group1;
m_VtxAttr.texCoord[1].Elements = m_group1.Tex1CoordElements;
m_VtxAttr.texCoord[1].Format = m_group1.Tex1CoordFormat;
m_VtxAttr.texCoord[1].Frac = m_group1.Tex1Frac;
m_VtxAttr.texCoord[2].Elements = m_group1.Tex2CoordElements;
m_VtxAttr.texCoord[2].Format = m_group1.Tex2CoordFormat;
m_VtxAttr.texCoord[2].Frac = m_group1.Tex2Frac;
m_VtxAttr.texCoord[3].Elements = m_group1.Tex3CoordElements;
m_VtxAttr.texCoord[3].Format = m_group1.Tex3CoordFormat;
m_VtxAttr.texCoord[3].Frac = m_group1.Tex3Frac;
m_VtxAttr.texCoord[4].Elements = m_group1.Tex4CoordElements;
m_VtxAttr.texCoord[4].Format = m_group1.Tex4CoordFormat;
};
void SetVAT_group2(u32 _group2)
{
if (m_group2.Hex == _group2)
return;
m_AttrDirty = true;
m_group2.Hex = _group2;
m_VtxAttr.texCoord[4].Frac = m_group2.Tex4Frac;
m_VtxAttr.texCoord[5].Elements = m_group2.Tex5CoordElements;
m_VtxAttr.texCoord[5].Format = m_group2.Tex5CoordFormat;
m_VtxAttr.texCoord[5].Frac = m_group2.Tex5Frac;
m_VtxAttr.texCoord[6].Elements = m_group2.Tex6CoordElements;
m_VtxAttr.texCoord[6].Format = m_group2.Tex6CoordFormat;
m_VtxAttr.texCoord[6].Frac = m_group2.Tex6Frac;
m_VtxAttr.texCoord[7].Elements = m_group2.Tex7CoordElements;
m_VtxAttr.texCoord[7].Format = m_group2.Tex7CoordFormat;
m_VtxAttr.texCoord[7].Frac = m_group2.Tex7Frac;
};
};
extern VertexLoader g_VertexLoaders[8];
extern DecodedVArray* varray;
#endif

140
Source/Plugins/Plugin_VideoDX9/Src/VertexLoaderManager.cpp
View File

@ -15,42 +15,142 @@
// Official SVN repository and contact information can be found at
// http://code.google.com/p/dolphin-emu/
#include "VideoCommon.h"
#include "VertexLoader.h"
#include "VertexManager.h"
#include <map>
DecodedVArray tempvarray;
#include "Statistics.h"
#include "VertexShaderManager.h"
#include "VertexLoader.h"
#include "VertexLoaderManager.h"
static int s_attr_dirty; // bitfield
static VertexLoader *g_VertexLoaders[8];
namespace VertexLoaderManager
{
typedef std::map<VertexLoaderUID, VertexLoader *> VertexLoaderMap;
static VertexLoaderMap g_VertexLoaderMap;
// TODO - change into array of pointers. Keep a map of all seen so far.
void Init()
{
tempvarray.Create(65536*3, 1, 8, 3, 2, 8);
MarkAllDirty();
for (int i = 0; i < 8; i++)
g_VertexLoaders[i] = NULL;
}
void Shutdown()
{
tempvarray.Destroy();
for (VertexLoaderMap::iterator iter = g_VertexLoaderMap.begin(); iter != g_VertexLoaderMap.end(); ++iter)
{
delete iter->second;
}
g_VertexLoaderMap.clear();
}
int GetVertexSize(int vat)
void AppendListToString(std::string *dest)
{
VertexLoader& vtxLoader = g_VertexLoaders[vat];
vtxLoader.Setup();
return vtxLoader.GetVertexSize();
for (VertexLoaderMap::iterator iter = g_VertexLoaderMap.begin(); iter != g_VertexLoaderMap.end(); ++iter)
{
iter->second->AppendToString(dest);
}
}
void RunVertices(int vat, int primitive, int num_vertices)
void MarkAllDirty()
{
tempvarray.Reset();
VertexLoader::SetVArray(&tempvarray);
VertexLoader& vtxLoader = g_VertexLoaders[vat];
vtxLoader.Setup();
vtxLoader.PrepareRun();
int vsize = vtxLoader.GetVertexSize();
vtxLoader.RunVertices(num_vertices);
VertexManager::AddVertices(primitive, num_vertices, &tempvarray);
s_attr_dirty = 0xff;
}
}
static void RefreshLoader(int vtx_attr_group)
{
if ((s_attr_dirty >> vtx_attr_group) & 1)
{
VertexLoaderUID uid;
uid.InitFromCurrentState(vtx_attr_group);
VertexLoaderMap::iterator iter = g_VertexLoaderMap.find(uid);
if (iter != g_VertexLoaderMap.end())
{
g_VertexLoaders[vtx_attr_group] = iter->second;
}
else
{
VertexLoader *loader = new VertexLoader(g_VtxDesc, g_VtxAttr[vtx_attr_group]);
g_VertexLoaderMap[uid] = loader;
g_VertexLoaders[vtx_attr_group] = loader;
INCSTAT(stats.numVertexLoaders);
}
}
s_attr_dirty &= ~(1 << vtx_attr_group);
}
void RunVertices(int vtx_attr_group, int primitive, int count)
{
if (!count)
return;
RefreshLoader(vtx_attr_group);
g_VertexLoaders[vtx_attr_group]->RunVertices(vtx_attr_group, primitive, count);
}
int GetVertexSize(int vtx_attr_group)
{
RefreshLoader(vtx_attr_group);
return g_VertexLoaders[vtx_attr_group]->GetVertexSize();
}
} // namespace
void LoadCPReg(u32 sub_cmd, u32 value)
{
switch (sub_cmd & 0xF0)
{
case 0x30:
VertexShaderManager::SetTexMatrixChangedA(value);
break;
case 0x40:
VertexShaderManager::SetTexMatrixChangedB(value);
break;
case 0x50:
g_VtxDesc.Hex &= ~0x1FFFF; // keep the Upper bits
g_VtxDesc.Hex |= value;
s_attr_dirty = 0xFF;
break;
case 0x60:
g_VtxDesc.Hex &= 0x1FFFF; // keep the lower 17Bits
g_VtxDesc.Hex |= (u64)value << 17;
s_attr_dirty = 0xFF;
break;
case 0x70:
_assert_((sub_cmd & 0x0F) < 8);
g_VtxAttr[sub_cmd & 7].g0.Hex = value;
s_attr_dirty |= 1 << (sub_cmd & 7);
break;
case 0x80:
_assert_((sub_cmd & 0x0F) < 8);
g_VtxAttr[sub_cmd & 7].g1.Hex = value;
s_attr_dirty |= 1 << (sub_cmd & 7);
break;
case 0x90:
_assert_((sub_cmd & 0x0F) < 8);
g_VtxAttr[sub_cmd & 7].g2.Hex = value;
s_attr_dirty |= 1 << (sub_cmd & 7);
break;
// Pointers to vertex arrays in GC RAM
case 0xA0:
arraybases[sub_cmd & 0xF] = value & 0xFFFFFFFF; // huh, why the mask?
break;
case 0xB0:
arraystrides[sub_cmd & 0xF] = value & 0xFF;
break;
}
}

226
Source/Plugins/Plugin_VideoDX9/Src/VertexLoader_Color.h
View File

@ -1,226 +0,0 @@
// 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/
#ifndef _VERTEXLOADERCOLOR_H
#define _VERTEXLOADERCOLOR_H
#include "LookUpTables.h"
#define RSHIFT 16
#define GSHIFT 8
#define BSHIFT 0
#define ASHIFT 24
extern DecodedVArray *varray;
extern int colIndex;
inline void _SetCol(u32 val)
{
varray->SetColor(colIndex, val);
colIndex++;
}
void _SetCol4444(u16 val)
{
u32 col = lut4to8[(val>>0)&0xF]<<ASHIFT;
col |= lut4to8[(val>>12)&0xF] <<RSHIFT;
col |= lut4to8[(val>>8)&0xF] <<GSHIFT;
col |= lut4to8[(val>>4)&0xF] <<BSHIFT;
_SetCol(col);
}
void _SetCol6666(u32 val)
{
u32 col = lut6to8[(val>>18)&0x3F] << RSHIFT;
col |= lut6to8[(val>>12)&0x3F] << GSHIFT;
col |= lut6to8[(val>>6)&0x3F] << BSHIFT;
col |= lut6to8[(val>>0)&0x3F] << ASHIFT;
_SetCol(col);
}
void _SetCol565(u16 val)
{
u32 col = lut5to8[(val>>11)&0x1f] << RSHIFT;
col |= lut6to8[(val>>5 )&0x3f] << GSHIFT;
col |= lut5to8[(val )&0x1f] << BSHIFT;
_SetCol(col | (0xFF<<ASHIFT));
}
//////////////////////////////////////////////////////////////////////////
//////////////////////////////////////////////////////////////////////////
inline u32 _Read24(u32 iAddress)
{
u32 col = Memory_Read_U8(iAddress) << RSHIFT; //should just get a pointer to main memory instead of going thru slow memhandler
col |= Memory_Read_U8(iAddress+1) << GSHIFT; //we can guarantee that it is reading from main memory
col |= Memory_Read_U8(iAddress+2) << BSHIFT;
return col | (0xFF<<ASHIFT);
}
inline u32 _Read32(u32 iAddress)
{
u32 col = Memory_Read_U8(iAddress) << RSHIFT; //should just get a pointer to main memory instead of going thru slow memhandler
col |= Memory_Read_U8(iAddress+1) << GSHIFT; //we can guarantee that it is reading from main memory
col |= Memory_Read_U8(iAddress+2) << BSHIFT;
col |= Memory_Read_U8(iAddress+3) << ASHIFT;
return col;
}
//////////////////////////////////////////////////////////////////////////
//////////////////////////////////////////////////////////////////////////
void LOADERDECL Color_ReadDirect_24b_888(const void *_p)
{
u32 col = DataReadU8()<<RSHIFT;
col |= DataReadU8()<<GSHIFT;
col |= DataReadU8()<<BSHIFT;
_SetCol(col | (0xFF<<ASHIFT));
}
void LOADERDECL Color_ReadDirect_32b_888x(const void *_p){
u32 col = DataReadU8()<<RSHIFT;
col |= DataReadU8()<<GSHIFT;
col |= DataReadU8()<<BSHIFT;
_SetCol(col | (0xFF<<ASHIFT));
DataReadU8();
}
void LOADERDECL Color_ReadDirect_16b_565(const void *_p)
{
_SetCol565(DataReadU16());
}
void LOADERDECL Color_ReadDirect_16b_4444(const void *_p)
{
_SetCol4444(DataReadU16());
}
void LOADERDECL Color_ReadDirect_24b_6666(const void *_p)
{
u32 val = DataReadU8()<<16;
val|=DataReadU8()<<8;
val|=DataReadU8();
_SetCol6666(val);
}
// F|RES: i am not 100 percent sure, but the colElements seems to be important for rendering only
// at least it fixes mario party 4
//
// if (colElements[colIndex])
// else
// col |= 0xFF<<ASHIFT;
//
void LOADERDECL Color_ReadDirect_32b_8888(const void *_p)
{
u32 col = DataReadU8()<<RSHIFT;
col |= DataReadU8()<<GSHIFT;
col |= DataReadU8()<<BSHIFT;
col |= DataReadU8()<<ASHIFT;
// "kill" the alpha
if (!colElements[colIndex])
col |= 0xFF << ASHIFT;
_SetCol(col);
}
//////////////////////////////////////////////////////////////////////////
//////////////////////////////////////////////////////////////////////////
//////////////////////////////////////////////////////////////////////////
void LOADERDECL Color_ReadIndex8_16b_565(const void *_p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
u16 val = Memory_Read_U16(iAddress);
_SetCol565(val);
}
void LOADERDECL Color_ReadIndex8_24b_888(const void *_p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
_SetCol(_Read24(iAddress));
}
void LOADERDECL Color_ReadIndex8_32b_888x(const void *_p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR]+colIndex);
_SetCol(_Read24(iAddress));
}
void LOADERDECL Color_ReadIndex8_16b_4444(const void *_p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
u16 val = Memory_Read_U16(iAddress);
_SetCol4444(val);
}
void LOADERDECL Color_ReadIndex8_24b_6666(const void *_p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
u32 val = Memory_Read_U8(iAddress+2) |
(Memory_Read_U8(iAddress+1)<<8) |
(Memory_Read_U8(iAddress)<<16);
_SetCol6666(val);
}
void LOADERDECL Color_ReadIndex8_32b_8888(const void *_p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
_SetCol(_Read32(iAddress));
}
//////////////////////////////////////////////////////////////////////////
//////////////////////////////////////////////////////////////////////////
//////////////////////////////////////////////////////////////////////////
void LOADERDECL Color_ReadIndex16_16b_565(const void *_p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
u16 val = Memory_Read_U16(iAddress);
_SetCol565(val);
}
void LOADERDECL Color_ReadIndex16_24b_888(const void *_p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
_SetCol(_Read24(iAddress));
}
void LOADERDECL Color_ReadIndex16_32b_888x(const void *_p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
_SetCol(_Read24(iAddress));
}
void LOADERDECL Color_ReadIndex16_16b_4444(const void *_p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
u16 val = Memory_Read_U16(iAddress);
_SetCol4444(val);
}
void LOADERDECL Color_ReadIndex16_24b_6666(const void *_p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
u32 val = Memory_Read_U8(iAddress+2) |
(Memory_Read_U8(iAddress+1)<<8) |
(Memory_Read_U8(iAddress)<<16);
_SetCol6666(val);
}
void LOADERDECL Color_ReadIndex16_32b_8888(const void *_p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_COLOR+colIndex] + (Index * arraystrides[ARRAY_COLOR+colIndex]);
_SetCol(_Read32(iAddress));
}
#endif

41
Source/Plugins/Plugin_VideoDX9/Src/VertexLoader_MtxIndex.h
View File

@ -1,41 +0,0 @@
// 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/
#pragma once
#include "Common.h"
#include "TransformEngine.h"
// ==============================================================================
// Direct
// ==============================================================================
void LOADERDECL PosMtx_ReadDirect_UByte(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
int index = DataReadU8() & 0x3f;
float *flipmem = (float *)xfmem;
varray->SetPosNrmIdx(index);
}
int s_texmtxread = 0, s_texmtxwrite = 0;
void LOADERDECL TexMtx_ReadDirect_UByte(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
int index = DataReadU8() & 0x3f;
varray->SetTcIdx(s_texmtxread++, index);
}

488
Source/Plugins/Plugin_VideoDX9/Src/VertexLoader_Normal.cpp
View File

@ -1,488 +0,0 @@
// 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 "main.h"
#include "Utils.h"
#include "DecodedVArray.h"
#include "VertexLoader.h"
#include "VertexLoader_Normal.h"
u8 VertexLoader_Normal::m_sizeTable[NUM_NRM_TYPE][NUM_NRM_FORMAT][NUM_NRM_ELEMENTS];
TPipelineFunction VertexLoader_Normal::m_funcTable[NUM_NRM_TYPE][NUM_NRM_FORMAT][NUM_NRM_ELEMENTS];
bool VertexLoader_Normal::index3;
// __________________________________________________________________________________________________
// Init
//
void VertexLoader_Normal::Init(void)
{
// size table
m_sizeTable[NRM_DIRECT][FORMAT_UBYTE] [NRM_NBT] = 3;
m_sizeTable[NRM_DIRECT][FORMAT_BYTE] [NRM_NBT] = 3;
m_sizeTable[NRM_DIRECT][FORMAT_USHORT][NRM_NBT] = 6;
m_sizeTable[NRM_DIRECT][FORMAT_SHORT] [NRM_NBT] = 6;
m_sizeTable[NRM_DIRECT][FORMAT_FLOAT] [NRM_NBT] = 12;
m_sizeTable[NRM_DIRECT][FORMAT_UBYTE] [NRM_NBT3] = 9;
m_sizeTable[NRM_DIRECT][FORMAT_BYTE] [NRM_NBT3] = 9;
m_sizeTable[NRM_DIRECT][FORMAT_USHORT][NRM_NBT3] = 18;
m_sizeTable[NRM_DIRECT][FORMAT_SHORT] [NRM_NBT3] = 18;
m_sizeTable[NRM_DIRECT][FORMAT_FLOAT] [NRM_NBT3] = 36;
m_sizeTable[NRM_INDEX8][FORMAT_UBYTE] [NRM_NBT] = 1;
m_sizeTable[NRM_INDEX8][FORMAT_BYTE] [NRM_NBT] = 1;
m_sizeTable[NRM_INDEX8][FORMAT_USHORT][NRM_NBT] = 1;
m_sizeTable[NRM_INDEX8][FORMAT_SHORT] [NRM_NBT] = 1;
m_sizeTable[NRM_INDEX8][FORMAT_FLOAT] [NRM_NBT] = 1;
m_sizeTable[NRM_INDEX8][FORMAT_UBYTE] [NRM_NBT3] = 3;
m_sizeTable[NRM_INDEX8][FORMAT_BYTE] [NRM_NBT3] = 3;
m_sizeTable[NRM_INDEX8][FORMAT_USHORT][NRM_NBT3] = 3;
m_sizeTable[NRM_INDEX8][FORMAT_SHORT] [NRM_NBT3] = 3;
m_sizeTable[NRM_INDEX8][FORMAT_FLOAT] [NRM_NBT3] = 3;
m_sizeTable[NRM_INDEX16][FORMAT_UBYTE] [NRM_NBT] = 2;
m_sizeTable[NRM_INDEX16][FORMAT_BYTE] [NRM_NBT] = 2;
m_sizeTable[NRM_INDEX16][FORMAT_USHORT][NRM_NBT] = 2;
m_sizeTable[NRM_INDEX16][FORMAT_SHORT] [NRM_NBT] = 2;
m_sizeTable[NRM_INDEX16][FORMAT_FLOAT] [NRM_NBT] = 2;
m_sizeTable[NRM_INDEX16][FORMAT_UBYTE] [NRM_NBT3] = 6;
m_sizeTable[NRM_INDEX16][FORMAT_BYTE] [NRM_NBT3] = 6;
m_sizeTable[NRM_INDEX16][FORMAT_USHORT][NRM_NBT3] = 6;
m_sizeTable[NRM_INDEX16][FORMAT_SHORT] [NRM_NBT3] = 6;
m_sizeTable[NRM_INDEX16][FORMAT_FLOAT] [NRM_NBT3] = 6;
// function table
m_funcTable[NRM_DIRECT][FORMAT_UBYTE] [NRM_NBT] = Normal_DirectByte; //HACK
m_funcTable[NRM_DIRECT][FORMAT_BYTE] [NRM_NBT] = Normal_DirectByte;
m_funcTable[NRM_DIRECT][FORMAT_USHORT][NRM_NBT] = Normal_DirectShort; //HACK
m_funcTable[NRM_DIRECT][FORMAT_SHORT] [NRM_NBT] = Normal_DirectShort;
m_funcTable[NRM_DIRECT][FORMAT_FLOAT] [NRM_NBT] = Normal_DirectFloat;
m_funcTable[NRM_DIRECT][FORMAT_UBYTE] [NRM_NBT3] = Normal_DirectByte3; //HACK
m_funcTable[NRM_DIRECT][FORMAT_BYTE] [NRM_NBT3] = Normal_DirectByte3;
m_funcTable[NRM_DIRECT][FORMAT_USHORT][NRM_NBT3] = Normal_DirectShort3; //HACK
m_funcTable[NRM_DIRECT][FORMAT_SHORT] [NRM_NBT3] = Normal_DirectShort3;
m_funcTable[NRM_DIRECT][FORMAT_FLOAT] [NRM_NBT3] = Normal_DirectFloat3;
m_funcTable[NRM_INDEX8][FORMAT_UBYTE] [NRM_NBT] = Normal_Index8_Byte; //HACK
m_funcTable[NRM_INDEX8][FORMAT_BYTE] [NRM_NBT] = Normal_Index8_Byte;
m_funcTable[NRM_INDEX8][FORMAT_USHORT][NRM_NBT] = Normal_Index8_Short; //HACK
m_funcTable[NRM_INDEX8][FORMAT_SHORT] [NRM_NBT] = Normal_Index8_Short;
m_funcTable[NRM_INDEX8][FORMAT_FLOAT] [NRM_NBT] = Normal_Index8_Float;
m_funcTable[NRM_INDEX8][FORMAT_UBYTE] [NRM_NBT3] = Normal_Index8_Byte3; //HACK
m_funcTable[NRM_INDEX8][FORMAT_BYTE] [NRM_NBT3] = Normal_Index8_Byte3;
m_funcTable[NRM_INDEX8][FORMAT_USHORT][NRM_NBT3] = Normal_Index8_Short3; //HACK
m_funcTable[NRM_INDEX8][FORMAT_SHORT] [NRM_NBT3] = Normal_Index8_Short3;
m_funcTable[NRM_INDEX8][FORMAT_FLOAT] [NRM_NBT3] = Normal_Index8_Float3;
m_funcTable[NRM_INDEX16][FORMAT_UBYTE] [NRM_NBT] = Normal_Index16_Byte; //HACK
m_funcTable[NRM_INDEX16][FORMAT_BYTE] [NRM_NBT] = Normal_Index16_Byte;
m_funcTable[NRM_INDEX16][FORMAT_USHORT][NRM_NBT] = Normal_Index16_Short; //HACK
m_funcTable[NRM_INDEX16][FORMAT_SHORT] [NRM_NBT] = Normal_Index16_Short;
m_funcTable[NRM_INDEX16][FORMAT_FLOAT] [NRM_NBT] = Normal_Index16_Float;
m_funcTable[NRM_INDEX16][FORMAT_UBYTE] [NRM_NBT3] = Normal_Index16_Byte3; //HACK
m_funcTable[NRM_INDEX16][FORMAT_BYTE] [NRM_NBT3] = Normal_Index16_Byte3;
m_funcTable[NRM_INDEX16][FORMAT_USHORT][NRM_NBT3] = Normal_Index16_Short3; //HACK
m_funcTable[NRM_INDEX16][FORMAT_SHORT] [NRM_NBT3] = Normal_Index16_Short3;
m_funcTable[NRM_INDEX16][FORMAT_FLOAT] [NRM_NBT3] = Normal_Index16_Float3;
}
// __________________________________________________________________________________________________
// GetSize
//
unsigned int
VertexLoader_Normal::GetSize(unsigned int _type, unsigned int _format, unsigned int _elements)
{
return m_sizeTable[_type][_format][_elements];
}
// __________________________________________________________________________________________________
// GetFunction
//
TPipelineFunction
VertexLoader_Normal::GetFunction(unsigned int _type, unsigned int _format, unsigned int _elements)
{
TPipelineFunction pFunc = m_funcTable[_type][_format][_elements];
return pFunc;
}
/////////////////////////////////////////////////////////////////////////////////////////////////////
//
// --- Direct ---
//
/////////////////////////////////////////////////////////////////////////////////////////////////////
// __________________________________________________________________________________________________
// Normal_DirectByte
//
void LOADERDECL
VertexLoader_Normal::Normal_DirectByte(const void* _p)
{
varray->SetNormalX(0, ((float)(signed char)DataReadU8()+0.5f) / 127.5f);
varray->SetNormalY(0, ((float)(signed char)DataReadU8()+0.5f) / 127.5f);
varray->SetNormalZ(0, ((float)(signed char)DataReadU8()+0.5f) / 127.5f);
}
// __________________________________________________________________________________________________
// Normal_DirectShort
//
void LOADERDECL
VertexLoader_Normal::Normal_DirectShort(const void* _p)
{
varray->SetNormalX(0, ((float)(signed short)DataReadU16()+0.5f) / 32767.5f);
varray->SetNormalY(0, ((float)(signed short)DataReadU16()+0.5f) / 32767.5f);
varray->SetNormalZ(0, ((float)(signed short)DataReadU16()+0.5f) / 32767.5f);
}
// __________________________________________________________________________________________________
// Normal_DirectFloat
//
void LOADERDECL
VertexLoader_Normal::Normal_DirectFloat(const void* _p)
{
varray->SetNormalX(0, DataReadF32());
varray->SetNormalY(0, DataReadF32());
varray->SetNormalZ(0, DataReadF32());
}
// __________________________________________________________________________________________________
// Normal_DirectByte3
//
void LOADERDECL
VertexLoader_Normal::Normal_DirectByte3(const void* _p)
{
for (int i=0; i<3; i++)
{
varray->SetNormalX(i, ((float)(signed char)DataReadU8()+0.5f) / 127.5f);
varray->SetNormalY(i, ((float)(signed char)DataReadU8()+0.5f) / 127.5f);
varray->SetNormalZ(i, ((float)(signed char)DataReadU8()+0.5f) / 127.5f);
}
}
// __________________________________________________________________________________________________
// Normal_DirectShort3
//
void LOADERDECL
VertexLoader_Normal::Normal_DirectShort3(const void* _p)
{
for (int i=0; i<3; i++)
{
varray->SetNormalX(i, ((float)(signed short)DataReadU16()+0.5f) / 32767.5f);
varray->SetNormalY(i, ((float)(signed short)DataReadU16()+0.5f) / 32767.5f);
varray->SetNormalZ(i, ((float)(signed short)DataReadU16()+0.5f) / 32767.5f);
}
}
// __________________________________________________________________________________________________
// Normal_DirectFloat3
//
void LOADERDECL
VertexLoader_Normal::Normal_DirectFloat3(const void* _p)
{
for (int i=0; i<3; i++)
{
varray->SetNormalX(i, DataReadF32());
varray->SetNormalY(i, DataReadF32());
varray->SetNormalZ(i, DataReadF32());
}
}
/////////////////////////////////////////////////////////////////////////////////////////////////////
//
// --- Index8 ---
//
/////////////////////////////////////////////////////////////////////////////////////////////////////
// __________________________________________________________________________________________________
// Normal_Index8_Byte
//
void LOADERDECL
VertexLoader_Normal::Normal_Index8_Byte(const void* _p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]);
varray->SetNormalX(0, ((float)(signed char)Memory_Read_U8(iAddress )+0.5f) / 127.5f);
varray->SetNormalY(0, ((float)(signed char)Memory_Read_U8(iAddress+1)+0.5f) / 127.5f);
varray->SetNormalZ(0, ((float)(signed char)Memory_Read_U8(iAddress+2)+0.5f) / 127.5f);
}
// __________________________________________________________________________________________________
// Normal_Index8_Short
//
void LOADERDECL
VertexLoader_Normal::Normal_Index8_Short(const void* _p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]);
varray->SetNormalX(0, ((float)(signed short)Memory_Read_U16(iAddress )+0.5f) / 32767.5f);
varray->SetNormalY(0, ((float)(signed short)Memory_Read_U16(iAddress+2)+0.5f) / 32767.5f);
varray->SetNormalZ(0, ((float)(signed short)Memory_Read_U16(iAddress+4)+0.5f) / 32767.5f);
}
// __________________________________________________________________________________________________
// Normal_Index8_Float
//
void LOADERDECL
VertexLoader_Normal::Normal_Index8_Float(const void* _p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]);
varray->SetNormalX(0, Memory_Read_Float(iAddress));
varray->SetNormalY(0, Memory_Read_Float(iAddress+4));
varray->SetNormalZ(0, Memory_Read_Float(iAddress+8));
}
// __________________________________________________________________________________________________
// Normal_Index8_Byte3
//
void LOADERDECL
VertexLoader_Normal::Normal_Index8_Byte3(const void* _p)
{
if (index3)
{
for (int i=0; i<3; i++)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 1*3*i;
varray->SetNormalX(i, ((float)(signed char)Memory_Read_U8(iAddress )+0.5f) / 127.5f);
varray->SetNormalY(i, ((float)(signed char)Memory_Read_U8(iAddress+1)+0.5f) / 127.5f);
varray->SetNormalZ(i, ((float)(signed char)Memory_Read_U8(iAddress+2)+0.5f) / 127.5f);
}
}
else
{
u8 Index = DataReadU8();
for (int i=0; i<3; i++)
{
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 1*3*i;
varray->SetNormalX(i, ((float)(signed char)Memory_Read_U8(iAddress )+0.5f) / 127.5f);
varray->SetNormalY(i, ((float)(signed char)Memory_Read_U8(iAddress+1)+0.5f) / 127.5f);
varray->SetNormalZ(i, ((float)(signed char)Memory_Read_U8(iAddress+2)+0.5f) / 127.5f);
}
}
}
// __________________________________________________________________________________________________
// Normal_Index8_Short3
//
void LOADERDECL
VertexLoader_Normal::Normal_Index8_Short3(const void* _p)
{
if (index3)
{
for (int i=0; i<3; i++)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 2*3*i;
varray->SetNormalX(i, ((float)(signed short)Memory_Read_U16(iAddress )+0.5f) / 32767.5f);
varray->SetNormalY(i, ((float)(signed short)Memory_Read_U16(iAddress+2)+0.5f) / 32767.5f);
varray->SetNormalZ(i, ((float)(signed short)Memory_Read_U16(iAddress+4)+0.5f) / 32767.5f);
}
}
else
{
u8 Index = DataReadU8();
for (int i=0; i<3; i++)
{
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 2*3*i;
varray->SetNormalX(i, ((float)(signed short)Memory_Read_U16(iAddress )+0.5f) / 32767.5f);
varray->SetNormalY(i, ((float)(signed short)Memory_Read_U16(iAddress+2)+0.5f) / 32767.5f);
varray->SetNormalZ(i, ((float)(signed short)Memory_Read_U16(iAddress+4)+0.5f) / 32767.5f);
}
}
}
// __________________________________________________________________________________________________
// Normal_Index8_Float3
//
void LOADERDECL
VertexLoader_Normal::Normal_Index8_Float3(const void* _p)
{
if (index3)
{
for (int i=0; i<3; i++)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 4*3*i;
varray->SetNormalX(i, Memory_Read_Float(iAddress));
varray->SetNormalY(i, Memory_Read_Float(iAddress+4));
varray->SetNormalZ(i, Memory_Read_Float(iAddress+8));
}
}
else
{
u8 Index = DataReadU8();
for (int i=0; i<3; i++)
{
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 4*3*i;
varray->SetNormalX(i, Memory_Read_Float(iAddress));
varray->SetNormalY(i, Memory_Read_Float(iAddress+4));
varray->SetNormalZ(i, Memory_Read_Float(iAddress+8));
}
}
}
/////////////////////////////////////////////////////////////////////////////////////////////////////
//
// --- Index16 ---
//
/////////////////////////////////////////////////////////////////////////////////////////////////////
// __________________________________________________________________________________________________
// Normal_Index16_Byte
//
void LOADERDECL
VertexLoader_Normal::Normal_Index16_Byte(const void* _p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]);
varray->SetNormalX(0, ((float)(signed char)Memory_Read_U8(iAddress )+0.5f) / 127.5f);
varray->SetNormalY(0, ((float)(signed char)Memory_Read_U8(iAddress+1)+0.5f) / 127.5f);
varray->SetNormalZ(0, ((float)(signed char)Memory_Read_U8(iAddress+2)+0.5f) / 127.5f);
}
// __________________________________________________________________________________________________
// Normal_Index16_Short
//
void LOADERDECL
VertexLoader_Normal::Normal_Index16_Short(const void* _p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]);
varray->SetNormalX(0, ((float)(signed short)Memory_Read_U16(iAddress )+0.5f) / 32767.5f);
varray->SetNormalY(0, ((float)(signed short)Memory_Read_U16(iAddress+2)+0.5f) / 32767.5f);
varray->SetNormalZ(0, ((float)(signed short)Memory_Read_U16(iAddress+4)+0.5f) / 32767.5f);
}
// __________________________________________________________________________________________________
// Normal_Index8_Float
//
void LOADERDECL
VertexLoader_Normal::Normal_Index16_Float(const void* _p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]);
varray->SetNormalX(0, Memory_Read_Float(iAddress));
varray->SetNormalY(0, Memory_Read_Float(iAddress+4));
varray->SetNormalZ(0, Memory_Read_Float(iAddress+8));
}
// __________________________________________________________________________________________________
// Normal_Index16_Byte3
//
void LOADERDECL
VertexLoader_Normal::Normal_Index16_Byte3(const void* _p)
{
if (index3)
{
for (int i=0; i<3; i++)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 1*3*i;
varray->SetNormalX(i, ((float)(signed char)Memory_Read_U8(iAddress )+0.5f) / 127.5f);
varray->SetNormalY(i, ((float)(signed char)Memory_Read_U8(iAddress+1)+0.5f) / 127.5f);
varray->SetNormalZ(i, ((float)(signed char)Memory_Read_U8(iAddress+2)+0.5f) / 127.5f);
}
}
else
{
u16 Index = DataReadU16();
for (int i=0; i<3; i++)
{
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 1*3*i;
varray->SetNormalX(i, ((float)(signed char)Memory_Read_U8(iAddress )+0.5f) / 127.5f);
varray->SetNormalY(i, ((float)(signed char)Memory_Read_U8(iAddress+1)+0.5f) / 127.5f);
varray->SetNormalZ(i, ((float)(signed char)Memory_Read_U8(iAddress+2)+0.5f) / 127.5f);
}
}
}
// __________________________________________________________________________________________________
// Normal_Index16_Short3
//
void LOADERDECL
VertexLoader_Normal::Normal_Index16_Short3(const void* _p)
{
if (index3)
{
for (int i=0; i<3; i++)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 2*3*i;
varray->SetNormalX(i, ((float)(signed short)Memory_Read_U16(iAddress )+0.5f) / 32767.5f);
varray->SetNormalY(i, ((float)(signed short)Memory_Read_U16(iAddress+2)+0.5f) / 32767.5f);
varray->SetNormalZ(i, ((float)(signed short)Memory_Read_U16(iAddress+4)+0.5f) / 32767.5f);
}
}
else
{
u16 Index = DataReadU16();
for (int i=0; i<3; i++)
{
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 2*3*i;
varray->SetNormalX(i, ((float)(signed short)Memory_Read_U16(iAddress )+0.5f) / 32767.5f);
varray->SetNormalY(i, ((float)(signed short)Memory_Read_U16(iAddress+2)+0.5f) / 32767.5f);
varray->SetNormalZ(i, ((float)(signed short)Memory_Read_U16(iAddress+4)+0.5f) / 32767.5f);
}
}
}
// __________________________________________________________________________________________________
// Normal_Index16_Float3
//
void LOADERDECL
VertexLoader_Normal::Normal_Index16_Float3(const void* _p)
{
if (index3)
{
for (int i=0; i<3; i++)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 4*3*i;
varray->SetNormalX(i, Memory_Read_Float(iAddress ));
varray->SetNormalY(i, Memory_Read_Float(iAddress+4));
varray->SetNormalZ(i, Memory_Read_Float(iAddress+8));
}
}
else
{
u16 Index = DataReadU16();
for (int i=0; i<3; i++)
{
u32 iAddress = arraybases[ARRAY_NORMAL] + (Index * arraystrides[ARRAY_NORMAL]) + 4*3*i;
varray->SetNormalX(i, Memory_Read_Float(iAddress ));
varray->SetNormalY(i, Memory_Read_Float(iAddress+4));
varray->SetNormalZ(i, Memory_Read_Float(iAddress+8));
}
}
}

90
Source/Plugins/Plugin_VideoDX9/Src/VertexLoader_Normal.h
View File

@ -1,90 +0,0 @@
// 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/
#pragma once
#include "CommonTypes.h"
class VertexLoader_Normal
{
public:
static bool index3;
// Init
static void Init(void);
// GetSize
static unsigned int GetSize(unsigned int _type, unsigned int _format, unsigned int _elements);
// GetFunction
static TPipelineFunction GetFunction(unsigned int _type, unsigned int _format, unsigned int _elements);
private:
enum ENormalType
{
NRM_NOT_PRESENT = 0,
NRM_DIRECT = 1,
NRM_INDEX8 = 2,
NRM_INDEX16 = 3,
NUM_NRM_TYPE
};
enum ENormalFormat
{
FORMAT_UBYTE = 0,
FORMAT_BYTE = 1,
FORMAT_USHORT = 2,
FORMAT_SHORT = 3,
FORMAT_FLOAT = 4,
NUM_NRM_FORMAT
};
enum ENormalElements
{
NRM_NBT = 0,
NRM_NBT3 = 1,
NUM_NRM_ELEMENTS
};
// tables
static u8 m_sizeTable[NUM_NRM_TYPE][NUM_NRM_FORMAT][NUM_NRM_ELEMENTS];
static TPipelineFunction m_funcTable[NUM_NRM_TYPE][NUM_NRM_FORMAT][NUM_NRM_ELEMENTS];
// direct
static void LOADERDECL Normal_DirectByte(const void* _p);
static void LOADERDECL Normal_DirectShort(const void* _p);
static void LOADERDECL Normal_DirectFloat(const void* _p);
static void LOADERDECL Normal_DirectByte3(const void* _p);
static void LOADERDECL Normal_DirectShort3(const void* _p);
static void LOADERDECL Normal_DirectFloat3(const void* _p);
// index8
static void LOADERDECL Normal_Index8_Byte(const void* _p);
static void LOADERDECL Normal_Index8_Short(const void* _p);
static void LOADERDECL Normal_Index8_Float(const void* _p);
static void LOADERDECL Normal_Index8_Byte3(const void* _p);
static void LOADERDECL Normal_Index8_Short3(const void* _p);
static void LOADERDECL Normal_Index8_Float3(const void* _p);
// index16
static void LOADERDECL Normal_Index16_Byte(const void* _p);
static void LOADERDECL Normal_Index16_Short(const void* _p);
static void LOADERDECL Normal_Index16_Float(const void* _p);
static void LOADERDECL Normal_Index16_Byte3(const void* _p);
static void LOADERDECL Normal_Index16_Short3(const void* _p);
static void LOADERDECL Normal_Index16_Float3(const void* _p);
};

213
Source/Plugins/Plugin_VideoDX9/Src/VertexLoader_Position.h
View File

@ -1,213 +0,0 @@
// 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/
#ifndef VERTEXHANDLER_POSITION_H
#define VERTEXHANDLER_POSITION_H
#include "Common.h"
#include "Utils.h"
// ==============================================================================
// Direct
// ==============================================================================
template<class T>
inline void LOADERDECL _ReadPos8Mem(int iAddress, TVtxAttr* pVtxAttr)
{
varray->SetPosX(((float)(T)Memory_Read_U8(iAddress)) * posScale);
varray->SetPosY(((float)(T)Memory_Read_U8(iAddress+1)) * posScale);
if (pVtxAttr->PosElements)
varray->SetPosZ(((float)(T)Memory_Read_U8(iAddress+2)) * posScale);
else
varray->SetPosZ(1.0);
}
template<class T>
inline void LOADERDECL _ReadPos16Mem(int iAddress, TVtxAttr* pVtxAttr)
{
varray->SetPosX(((float)(T)Memory_Read_U16(iAddress)) * posScale);
varray->SetPosY(((float)(T)Memory_Read_U16(iAddress+2)) * posScale);
if (pVtxAttr->PosElements)
varray->SetPosZ(((float)(T)Memory_Read_U16(iAddress+4)) * posScale);
else
varray->SetPosZ(1.0);
}
void LOADERDECL _ReadPosFloatMem(int iAddress, TVtxAttr* pVtxAttr)
{
u32 uTemp;
uTemp = Memory_Read_U32(iAddress );
varray->SetPosX(*(float*)&uTemp);
uTemp = Memory_Read_U32(iAddress+4);
varray->SetPosY(*(float*)&uTemp);
if (pVtxAttr->PosElements)
{
uTemp = Memory_Read_U32(iAddress+8);
varray->SetPosZ(*(float*)&uTemp);
}
else
varray->SetPosZ(1.0);
}
void LOADERDECL Pos_ReadDirect_UByte(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
varray->SetPosX((float)DataReadU8() * posScale);
varray->SetPosY((float)DataReadU8() * posScale);
if (pVtxAttr->PosElements)
varray->SetPosZ((float)DataReadU8() * posScale);
else
varray->SetPosZ(1.0);
}
void LOADERDECL Pos_ReadDirect_Byte(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
varray->SetPosX((float)(s8)DataReadU8() * posScale);
varray->SetPosY((float)(s8)DataReadU8() * posScale);
if (pVtxAttr->PosElements)
varray->SetPosZ((float)(s8)DataReadU8() * posScale);
else
varray->SetPosZ(1.0);
}
void LOADERDECL Pos_ReadDirect_UShort(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
varray->SetPosX((float)DataReadU16() * posScale);
varray->SetPosY((float)DataReadU16() * posScale);
if (pVtxAttr->PosElements)
varray->SetPosZ((float)DataReadU16() * posScale);
else
varray->SetPosZ(1.0);
}
void LOADERDECL Pos_ReadDirect_Short(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
varray->SetPosX((float)(s16)DataReadU16() * posScale);
varray->SetPosY((float)(s16)DataReadU16() * posScale);
if (pVtxAttr->PosElements)
varray->SetPosZ((float)(s16)DataReadU16() * posScale);
else
varray->SetPosZ(1.0);
}
void LOADERDECL Pos_ReadDirect_Float(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
varray->SetPosX(DataReadF32());
varray->SetPosY(DataReadF32());
if (pVtxAttr->PosElements)
varray->SetPosZ(DataReadF32());
else
varray->SetPosZ(1.0);
}
// ==============================================================================
// Index 8
// ==============================================================================
void LOADERDECL Pos_ReadIndex8_UByte(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_POSITION] + (Index * arraystrides[ARRAY_POSITION]);
_ReadPos8Mem<u8>(iAddress, pVtxAttr);
}
void LOADERDECL Pos_ReadIndex8_Byte(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_POSITION] + (Index * arraystrides[ARRAY_POSITION]);
_ReadPos8Mem<s8>(iAddress, pVtxAttr);
}
void LOADERDECL Pos_ReadIndex8_UShort(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_POSITION] + (Index * arraystrides[ARRAY_POSITION]);
_ReadPos16Mem<u16>(iAddress, pVtxAttr);
}
void LOADERDECL Pos_ReadIndex8_Short(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_POSITION] + (Index * arraystrides[ARRAY_POSITION]);
_ReadPos16Mem<s16>(iAddress, pVtxAttr);
}
void LOADERDECL Pos_ReadIndex8_Float(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_POSITION] + (Index * arraystrides[ARRAY_POSITION]);
_ReadPosFloatMem(iAddress, pVtxAttr);
}
// ==============================================================================
// Index 16
// ==============================================================================
void LOADERDECL Pos_ReadIndex16_UByte(const void* _p){
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_POSITION] + (Index * arraystrides[ARRAY_POSITION]);
_ReadPos8Mem<u8>(iAddress, pVtxAttr);
}
void LOADERDECL Pos_ReadIndex16_Byte(const void* _p){
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_POSITION] + (Index * arraystrides[ARRAY_POSITION]);
_ReadPos8Mem<s8>(iAddress, pVtxAttr);
}
void LOADERDECL Pos_ReadIndex16_UShort(const void* _p){
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_POSITION] + (Index * arraystrides[ARRAY_POSITION]);
_ReadPos16Mem<u16>(iAddress, pVtxAttr);
}
void LOADERDECL Pos_ReadIndex16_Short(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_POSITION] + (Index * arraystrides[ARRAY_POSITION]);
_ReadPos16Mem<s16>(iAddress, pVtxAttr);
}
void LOADERDECL Pos_ReadIndex16_Float(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_POSITION] + (Index * arraystrides[ARRAY_POSITION]);
_ReadPosFloatMem(iAddress, pVtxAttr);
}
#endif

202
Source/Plugins/Plugin_VideoDX9/Src/VertexLoader_TextCoord.h
View File

@ -1,202 +0,0 @@
// 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/
#ifndef VERTEXHANDLER_TEXCOORD_H
#define VERTEXHANDLER_TEXCOORD_H
extern int tcIndex;
void LOADERDECL TexCoord_ReadDirect_UByte(const void* _p)
{
varray->SetU(tcIndex, DataReadU8() * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
varray->SetV(tcIndex, DataReadU8() * tcScaleV[tcIndex]);
tcIndex++;
}
void LOADERDECL TexCoord_ReadDirect_Byte(const void* _p)
{
varray->SetU(tcIndex, (s8)DataReadU8() * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
varray->SetV(tcIndex, (s8)DataReadU8() * tcScaleV[tcIndex]);
tcIndex++;
}
void LOADERDECL TexCoord_ReadDirect_UShort(const void* _p)
{
varray->SetU(tcIndex, DataReadU16() * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
varray->SetV(tcIndex, DataReadU16() * tcScaleV[tcIndex]);
tcIndex++;
}
void LOADERDECL TexCoord_ReadDirect_Short(const void* _p)
{
varray->SetU(tcIndex, (s16)DataReadU16() * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
varray->SetV(tcIndex, (s16)DataReadU16() * tcScaleV[tcIndex]);
tcIndex++;
}
void LOADERDECL TexCoord_ReadDirect_Float(const void* _p)
{
varray->SetU(tcIndex, DataReadF32() * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
varray->SetV(tcIndex, DataReadF32() * tcScaleV[tcIndex]);
tcIndex++;
}
// ==================================================================================
void LOADERDECL TexCoord_ReadIndex8_UByte(const void* _p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_TEXCOORD0+tcIndex] + (Index * arraystrides[ARRAY_TEXCOORD0+tcIndex]);
varray->SetU(tcIndex, (float)(u8)Memory_Read_U8(iAddress) * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
varray->SetV(tcIndex, (float)(u8)Memory_Read_U8(iAddress+1) * tcScaleV[tcIndex]);
tcIndex++;
}
void LOADERDECL TexCoord_ReadIndex8_Byte(const void* _p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_TEXCOORD0+tcIndex] + (Index * arraystrides[ARRAY_TEXCOORD0+tcIndex]);
varray->SetU(tcIndex, (float)(s8)Memory_Read_U8(iAddress) * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
varray->SetV(tcIndex, (float)(s8)Memory_Read_U8(iAddress+1) * tcScaleV[tcIndex]);
tcIndex++;
}
void LOADERDECL TexCoord_ReadIndex8_UShort(const void* _p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_TEXCOORD0+tcIndex] + (Index * arraystrides[ARRAY_TEXCOORD0+tcIndex]);
varray->SetU(tcIndex, (float)(u16)Memory_Read_U16(iAddress) * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
varray->SetV(tcIndex, (float)(u16)Memory_Read_U16(iAddress+2) * tcScaleV[tcIndex]);
tcIndex++;
}
void LOADERDECL TexCoord_ReadIndex8_Short(const void* _p)
{
u8 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_TEXCOORD0+tcIndex] + (Index * arraystrides[ARRAY_TEXCOORD0+tcIndex]);
varray->SetU(tcIndex, (float)(s16)Memory_Read_U16(iAddress) * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
varray->SetV(tcIndex, (float)(s16)Memory_Read_U16(iAddress+2) * tcScaleV[tcIndex]);
tcIndex++;
}
void LOADERDECL TexCoord_ReadIndex8_Float(const void* _p)
{
u16 Index = DataReadU8();
u32 iAddress = arraybases[ARRAY_TEXCOORD0+tcIndex] + (Index * arraystrides[ARRAY_TEXCOORD0+tcIndex]);
u32 uTemp;
uTemp = Memory_Read_U32(iAddress );
varray->SetU(tcIndex, *(float*)&uTemp * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
{
uTemp = Memory_Read_U32(iAddress+4);
varray->SetV(tcIndex, *(float*)&uTemp * tcScaleV[tcIndex]);
}
tcIndex++;
}
// ==================================================================================
void LOADERDECL TexCoord_ReadIndex16_UByte(const void* _p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_TEXCOORD0+tcIndex] + (Index * arraystrides[ARRAY_TEXCOORD0+tcIndex]);
u32 uTemp;
uTemp = (u8)Memory_Read_U8(iAddress );
varray->SetU(tcIndex, float(uTemp)* tcScaleU[tcIndex]);
if (tcElements[tcIndex])
{
uTemp = (u8)Memory_Read_U8(iAddress+1);
varray->SetV(tcIndex, float(uTemp)* tcScaleV[tcIndex]);
}
tcIndex++;
}
void LOADERDECL TexCoord_ReadIndex16_Byte(const void* _p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_TEXCOORD0+tcIndex] + (Index * arraystrides[ARRAY_TEXCOORD0+tcIndex]);
u32 uTemp;
uTemp = (s8)Memory_Read_U8(iAddress );
varray->SetU(tcIndex, float(uTemp)* tcScaleU[tcIndex]);
if (tcElements[tcIndex])
{
uTemp = (s8)Memory_Read_U8(iAddress+1);
varray->SetV(tcIndex, float(uTemp)* tcScaleV[tcIndex]);
}
tcIndex++;
}
void LOADERDECL TexCoord_ReadIndex16_UShort(const void* _p)
{
TVtxAttr* pVtxAttr = (TVtxAttr*)_p;
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_TEXCOORD0+tcIndex] + (Index * arraystrides[ARRAY_TEXCOORD0+tcIndex]);
float uTemp;
uTemp = (float)(u16)Memory_Read_U16(iAddress );
varray->SetU(tcIndex, uTemp * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
{
uTemp = (float)(u16)Memory_Read_U16(iAddress+2);
varray->SetV(tcIndex, uTemp * tcScaleV[tcIndex]);
}
tcIndex++;
}
void LOADERDECL TexCoord_ReadIndex16_Short(const void* _p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_TEXCOORD0+tcIndex] + (Index * arraystrides[ARRAY_TEXCOORD0+tcIndex]);
float uTemp;
uTemp = (float)(s16)Memory_Read_U16(iAddress );
varray->SetU(tcIndex, uTemp * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
{
uTemp = (float)(s16)Memory_Read_U16(iAddress+2);
varray->SetV(tcIndex, uTemp * tcScaleV[tcIndex]);
}
tcIndex++;
}
void LOADERDECL TexCoord_ReadIndex16_Float(const void* _p)
{
u16 Index = DataReadU16();
u32 iAddress = arraybases[ARRAY_TEXCOORD0+tcIndex] + (Index * arraystrides[ARRAY_TEXCOORD0+tcIndex]);
u32 uTemp;
uTemp = Memory_Read_U32(iAddress );
varray->SetU(tcIndex, *(float*)&uTemp * tcScaleU[tcIndex]);
if (tcElements[tcIndex])
{
uTemp = Memory_Read_U32(iAddress+4);
varray->SetV(tcIndex, *(float*)&uTemp * tcScaleV[tcIndex]);
}
tcIndex++;
}
#endif

114
Source/Plugins/Plugin_VideoDX9/Src/VertexManager.cpp
View File

@ -23,28 +23,32 @@
#include "Profiler.h"
#include "VertexManager.h"
#include "OpcodeDecoding.h"
#include "TransformEngine.h"
#include "IndexGenerator.h"
#include "BPStructs.h"
#include "XFStructs.h"
#include "VertexShaderManager.h"
#include "VertexShaderCache.h"
#include "PixelShaderManager.h"
#include "PixelShaderCache.h"
#include "Utils.h"
#include "NativeVertexFormat.h"
#include "NativeVertexWriter.h"
#include "BPStructs.h"
#include "XFStructs.h"
using namespace D3D;
// internal state for loading vertices
extern NativeVertexFormat *g_nativeVertexFmt;
namespace VertexManager
{
enum Collection
{
C_NOTHING=0,
C_TRIANGLES=1,
C_LINES=2,
C_POINTS=3
C_NOTHING,
C_TRIANGLES,
C_LINES,
C_POINTS,
};
static IndexGenerator indexGen;
@ -52,29 +56,12 @@ static Collection collection;
static LPDIRECT3DVERTEXDECLARATION9 vDecl;
static D3DVertex *fakeVBuffer;
static u16 *fakeIBuffer;
static u8 *fakeVBuffer; // format undefined - NativeVertexFormat takes care of the declaration.
static u16 *fakeIBuffer; // These are just straightforward 16-bit indices.
#define MAXVBUFFERSIZE 65536*3
#define MAXIBUFFERSIZE 65536*3
const D3DVERTEXELEMENT9 decl[] =
{
{ 0, 0, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_POSITION, 0 },
{ 0, 12, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_NORMAL, 0 },
{ 0, 24, D3DDECLTYPE_D3DCOLOR, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_COLOR, 0 },
{ 0, 28, D3DDECLTYPE_D3DCOLOR, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_COLOR, 1 },
{ 0, 32+12*0, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_TEXCOORD, 0 },
{ 0, 32+12*1, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_TEXCOORD, 1 },
{ 0, 32+12*2, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_TEXCOORD, 2 },
{ 0, 32+12*3, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_TEXCOORD, 3 },
{ 0, 32+12*4, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_TEXCOORD, 4 },
{ 0, 32+12*5, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_TEXCOORD, 5 },
{ 0, 32+12*6, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_TEXCOORD, 6 },
{ 0, 32+12*7, D3DDECLTYPE_FLOAT3, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_TEXCOORD, 7 },
D3DDECL_END()
};
const Collection collectionTypeLUT[8] =
{
C_TRIANGLES,//quads
@ -87,18 +74,16 @@ const Collection collectionTypeLUT[8] =
C_POINTS //guess :P
};
D3DVertex *vbufferwrite;
void CreateDeviceObjects();
void DestroyDeviceObjects();
bool Init()
{
collection = C_NOTHING;
fakeVBuffer = new D3DVertex[65536];
fakeIBuffer = new u16[65536];
fakeVBuffer = new u8[MAXVBUFFERSIZE];
fakeIBuffer = new u16[MAXIBUFFERSIZE];
CreateDeviceObjects();
VertexManager::s_pCurBufferPointer = fakeVBuffer;
return true;
}
@ -111,30 +96,20 @@ void Shutdown()
void CreateDeviceObjects()
{
HRESULT hr;
if (FAILED(hr = D3D::dev->CreateVertexDeclaration(decl, &vDecl)))
{
MessageBox(0,"Failed to create vertex declaration","sdfsd",0);
return;
}
}
void BeginFrame()
{
D3D::dev->SetVertexDeclaration(vDecl);
//D3D::dev->SetStreamSource(0,vBuffer,0,sizeof(D3DVertex));
}
void DestroyDeviceObjects()
{
if (vDecl)
vDecl->Release();
vDecl = 0;
}
void AddIndices(int _primitive, int _numVertices)
{
switch (_primitive) {
switch (_primitive)
{
case GX_DRAW_QUADS: indexGen.AddQuads(_numVertices); return;
case GX_DRAW_TRIANGLES: indexGen.AddList(_numVertices); return;
case GX_DRAW_TRIANGLE_STRIP: indexGen.AddStrip(_numVertices); return;
@ -145,7 +120,12 @@ void AddIndices(int _primitive, int _numVertices)
}
}
void AddVertices(int _primitive, int _numVertices, const DecodedVArray *varray)
int GetRemainingSize()
{
return 60000;
}
void AddVertices(int _primitive, int _numVertices)
{
if (_numVertices <= 0) //This check is pretty stupid...
return;
@ -168,32 +148,24 @@ void AddVertices(int _primitive, int _numVertices, const DecodedVArray *varray)
Flush();
collection = type;
u16 *ptr = 0;
if (type != C_POINTS)
{
ptr = fakeIBuffer;
indexGen.Start((unsigned short*)ptr);
AddIndices(_primitive,_numVertices);
AddIndices(_primitive, _numVertices);
}
vbufferwrite = fakeVBuffer;
if (_numVertices >= MAXVBUFFERSIZE)
MessageBox(NULL, "To much vertices for the buffer", "Video.DLL", MB_OK);
CTransformEngine::TransformVertices(_numVertices, varray, vbufferwrite);
MessageBox(NULL, "Too many vertices for the buffer", "Dolphin DX9 Video Plugin", MB_OK);
}
else //We are collecting the right type, keep going
{
_assert_msg_(vbufferwrite!=0, "collecting: vbufferwrite == 0!","WTF");
_assert_msg_(vbufferwrite != 0, "collecting: vbufferwrite == 0!","WTF");
INCSTAT(stats.thisFrame.numPrimitiveJoins);
//Success, keep adding to unlocked buffer
int last = indexGen.GetNumVerts();
AddIndices(_primitive, _numVertices);
if (_numVertices >= MAXVBUFFERSIZE)
MessageBox(NULL, "Too many vertices for the buffer", "Video.DLL", MB_OK);
CTransformEngine::TransformVertices(_numVertices, varray, vbufferwrite + last);
MessageBox(NULL, "Too many vertices for the buffer", "Dolphin DX9 Video Plugin", MB_OK);
}
}
@ -206,37 +178,45 @@ const D3DPRIMITIVETYPE pts[3] =
void Flush()
{
DVSTARTPROFILE();
if (collection != C_NOTHING)
DVSTARTPROFILE();
if (collection != C_NOTHING)
{
ActivateTextures();
int numVertices = indexGen.GetNumVerts();
if (numVertices != 0)
if (numVertices)
{
PShaderCache::SetShader(); // TODO(ector): only do this if shader has changed
VShaderCache::SetShader(); // TODO(ector): only do this if shader has changed
PixelShaderCache::SetShader(); // TODO(ector): only do this if shader has changed
VertexShaderCache::SetShader(g_nativeVertexFmt->m_components); // TODO(ector): only do this if shader has changed
// set global constants
VertexShaderManager::SetConstants(false, false);
PixelShaderManager::SetConstants();
int stride = g_nativeVertexFmt->GetVertexStride();
g_nativeVertexFmt->SetupVertexPointers();
if (collection != C_POINTS)
{
int numPrimitives = indexGen.GetNumPrims();
D3D::dev->DrawIndexedPrimitiveUP(pts[(int)collection],
0,
0,
numVertices,
numPrimitives,
fakeIBuffer,
D3DFMT_INDEX16,
fakeVBuffer,
sizeof(D3DVertex));
stride);
}
else
{
D3D::dev->SetIndices(0);
D3D::dev->DrawPrimitiveUP(D3DPT_POINTLIST, numVertices, fakeVBuffer, sizeof(D3DVertex));
D3D::dev->DrawPrimitiveUP(D3DPT_POINTLIST, numVertices, fakeVBuffer, stride);
}
INCSTAT(stats.thisFrame.numDrawCalls);
}
collection = C_NOTHING;
VertexManager::s_pCurBufferPointer = fakeVBuffer;
}
}
} // namespace
} // namespace

15
Source/Plugins/Plugin_VideoDX9/Src/VertexManager.h
View File

@ -20,19 +20,6 @@
#include "CPMemory.h"
#include "VertexLoader.h"
#include "DecodedVArray.h"
struct UV
{
float u, v, w;
};
struct D3DVertex {
Vec3 pos;
Vec3 normal;
u32 colors[2];
UV uv[8];
};
namespace VertexManager
{
@ -42,7 +29,7 @@ void Shutdown();
void BeginFrame();
void AddVertices(int _primitive, int _numVertices, const DecodedVArray *varray);
void AddVertices(int _primitive, int _numVertices);
void Flush();
} // namespace

141
Source/Plugins/Plugin_VideoDX9/Src/VertexShader.cpp
View File

@ -1,141 +0,0 @@
// 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 "stdafx.h"
#include "D3DShader.h"
#include "VertexShader.h"
#include "BPStructs.h"
static const char *genericVS = "// Generic Vertex Shader\
\n\
struct VS_INPUT {\n\
float4 pos : POSITION;\n\
float3 normal : NORMAL;\n\
float4 colors[2] : COLOR0;\n\
float4 uv[5] : TEXCOORD0;\n\
};\n\
\n\
struct VS_OUTPUT {\n\
float4 pos : POSITION;\n\
float4 colors[2] : COLOR0;\n\
//numtexgen\n\
float4 uv[5] : TEXCOORD0;\n\
};\n\
\n\
uniform matrix matWorldViewProj : register(c0);\n\
\n\
VS_OUTPUT main(const VS_INPUT input)\n\
{\n\
VS_OUTPUT output;\n\
\n\
output.pos = mul(input.pos, matWorldViewProj);\n\
// texgen\n\
for (int i=0; i<5; i++)\n output.uv[i] = input.uv[i];\n\
\n\
for (int i=0; i<2; i++)\n output.colors[i] = input.colors[i];\n\
return output;\n\
}\0";
const char *GenerateVertexShader() {
return genericVS;
}
/*
char text2[65536];
#define WRITE p+=sprintf
void WriteTexgen(char *&p, int n);
const char *GenerateVertexShader()
{
int numColors = 2;
int numUV = 8;
int numTexgen = bpmem.genMode.numtexgens;
int numNormals = 3;
bool fogEnable = false;
bool hasNormal = true;
char *p = text2;
WRITE(p,"//Vertex Shader\n");
WRITE(p,"//%i uv->%i texgens, %i colors\n",numUV,numTexgen,numColors);
WRITE(p,"\n");
WRITE(p,"struct VS_INPUT {\n");
WRITE(p," float4 pos : POSITION;\n");
WRITE(p," float3 normal : NORMAL;\n");
if (numColors)
WRITE(p," float4 colors[%i] : COLOR0;\n",numColors);
if (numUV)
WRITE(p," float3 uv[%i] : TEXCOORD0;\n",numUV);
WRITE(p,"};\n");
WRITE(p,"\n");
WRITE(p,"struct VS_OUTPUT {\n");
WRITE(p," float4 pos : POSITION;\n");
WRITE(p," float4 colors[%i] : COLOR0;\n",numColors);
if (numTexgen)
WRITE(p," float4 uv[%i] : TEXCOORD0;\n",numTexgen);
if (fogEnable)
WRITE(p," float fog : FOG;\n",numTexgen);
WRITE(p,"};\n");
WRITE(p,"\n");
WRITE(p,"uniform matrix matWorldViewProj : register(c0);\n");
WRITE(p,"\n");
WRITE(p,"VS_OUTPUT main(const VS_INPUT input)\n");
WRITE(p,"{\n");
WRITE(p," VS_OUTPUT output;");
WRITE(p,"\n");
WRITE(p," output.pos = mul(matWorldViewProj, input.pos);\n");
for (int i = 0; i < (int)bpmem.genMode.numtexgens; i++)
{
//build the equation for this stage
WriteTexgen(p,i);
}
WRITE(p," for (int i=0; i<2; i++)\n output.colors[i] = input.colors[i];\n");
//WRITE(p," output.fog = 0.0f;");
WRITE(p,"return output;\n");
WRITE(p,"}\n");
WRITE(p,"\0");
// MessageBox(0,text2,0,0);
return text2;
}
/*
* xform->vertexshader ideas
*//*
void WriteTexgen(char *&p, int n)
{
WRITE(p," output.uv[%i] = float4(input.uv[%i].xy,0,input.uv[%i].z);\n",n,n,n);
}
void WriteLight(int color, int component)
{
} */

50
Source/Plugins/Plugin_VideoDX9/Src/VertexShader.h
View File

@ -1,50 +0,0 @@
// 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/
#ifndef _VERTEXSHADER_H
#define _VERTEXSHADER_H
#include "Common.h"
const char *GenerateVertexShader();
// shader variables
#define I_POSNORMALMATRIX "cpnmtx"
#define I_PROJECTION "cproj"
#define I_MATERIALS "cmtrl"
#define I_LIGHTS "clights"
#define I_TEXMATRICES "ctexmtx"
#define I_TRANSFORMMATRICES "ctrmtx"
#define I_NORMALMATRICES "cnmtx"
#define I_POSTTRANSFORMMATRICES "cpostmtx"
#define I_FOGPARAMS "cfog"
#define C_POSNORMALMATRIX 0
#define C_PROJECTION (C_POSNORMALMATRIX+6)
#define C_MATERIALS (C_PROJECTION+4)
#define C_LIGHTS (C_MATERIALS+4)
#define C_TEXMATRICES (C_LIGHTS+40)
#define C_TRANSFORMMATRICES (C_TEXMATRICES+24)
#define C_NORMALMATRICES (C_TRANSFORMMATRICES+64)
#define C_POSTTRANSFORMMATRICES (C_NORMALMATRICES+32)
#define C_FOGPARAMS (C_POSTTRANSFORMMATRICES+64)
#define PS_CONST_COLORS 0
#define PS_CONST_KCOLORS 4
#define PS_CONST_CONSTALPHA 8
#endif

45
Source/Plugins/Plugin_VideoDX9/Src/VertexShaderCache.cpp
View File

@ -18,6 +18,7 @@
#include <map>
#include "D3DBase.h"
#include "D3DShader.h"
#include "Statistics.h"
#include "Utils.h"
#include "Profiler.h"
@ -26,15 +27,27 @@
#include "BPMemory.h"
#include "XFMemory.h"
VShaderCache::VSCache VShaderCache::vshaders;
VertexShaderCache::VSCache VertexShaderCache::vshaders;
void VShaderCache::Init()
void SetVSConstant4f(int const_number, float f1, float f2, float f3, float f4)
{
const float f[4] = {f1, f2, f3, f4};
D3D::dev->SetVertexShaderConstantF(const_number, f, 1);
}
void SetVSConstant4fv(int const_number, const float *f)
{
D3D::dev->SetVertexShaderConstantF(const_number, f, 1);
}
void VertexShaderCache::Init()
{
}
void VShaderCache::Shutdown()
void VertexShaderCache::Shutdown()
{
VSCache::iterator iter = vshaders.begin();
for (; iter != vshaders.end(); iter++)
@ -43,28 +56,24 @@ void VShaderCache::Shutdown()
}
void VShaderCache::SetShader()
void VertexShaderCache::SetShader(u32 components)
{
static LPDIRECT3DVERTEXSHADER9 shader = NULL;
if (D3D::GetShaderVersion() < 2)
return; // we are screwed
if (shader) {
//D3D::dev->SetVertexShader(shader);
return;
}
static LPDIRECT3DVERTEXSHADER9 lastShader = 0;
DVSTARTPROFILE();
u32 currentHash = 0x1337; // GetCurrentTEV();
VERTEXSHADERUID uid;
GetVertexShaderId(uid, components, false);
VSCache::iterator iter;
iter = vshaders.find(currentHash);
iter = vshaders.find(uid);
if (iter != vshaders.end())
{
iter->second.frameCount=frameCount;
iter->second.frameCount = frameCount;
VSCacheEntry &entry = iter->second;
if (!lastShader || entry.shader != lastShader)
{
@ -74,15 +83,15 @@ void VShaderCache::SetShader()
return;
}
const char *code = GenerateVertexShader();
shader = D3D::CompileVShader(code, int(strlen(code)));
const char *code = GenerateVertexShader(components, false);
shader = D3D::CompileVertexShader(code, (int)strlen(code));
if (shader)
{
//Make an entry in the table
VSCacheEntry entry;
entry.shader = shader;
entry.frameCount=frameCount;
vshaders[currentHash] = entry;
entry.frameCount = frameCount;
vshaders[uid] = entry;
}
D3D::dev->SetVertexShader(shader);
@ -91,7 +100,7 @@ void VShaderCache::SetShader()
SETSTAT(stats.numVertexShadersAlive, (int)vshaders.size());
}
void VShaderCache::Cleanup()
void VertexShaderCache::Cleanup()
{
for (VSCache::iterator iter = vshaders.begin(); iter != vshaders.end();)
{
@ -107,4 +116,4 @@ void VShaderCache::Cleanup()
}
}
SETSTAT(stats.numVertexShadersAlive, (int)vshaders.size());
}
}

10
Source/Plugins/Plugin_VideoDX9/Src/VertexShaderCache.h
View File

@ -22,10 +22,10 @@
#include <map>
#include "PixelShader.h"
#include "VertexShader.h"
#include "D3DBase.h"
#include "VertexShaderGen.h"
class VShaderCache
class VertexShaderCache
{
struct VSCacheEntry
{
@ -43,7 +43,7 @@ class VShaderCache
}
};
typedef std::map<u32, VSCacheEntry> VSCache;
typedef std::map<VERTEXSHADERUID, VSCacheEntry> VSCache;
static VSCache vshaders;
@ -51,7 +51,7 @@ public:
static void Init();
static void Cleanup();
static void Shutdown();
static void SetShader();
static void SetShader(u32 components);
};
#endif // _VERTEXSHADERCACHE_H

1
Source/Plugins/Plugin_VideoDX9/Src/main.cpp
View File

@ -31,7 +31,6 @@
#include "BPStructs.h"
#include "VertexManager.h"
#include "VertexLoaderManager.h"
#include "TransformEngine.h"
#include "DlgSettings.h"
#include "D3DPostprocess.h"
#include "D3DTexture.h"

2
Source/Plugins/Plugin_VideoDX9/Src/main.h
View File

@ -23,6 +23,4 @@
extern SVideoInitialize g_VideoInitialize;
void DebugLog(const char* _fmt, ...);
#endif

8
Source/Plugins/Plugin_VideoOGL/Plugin_VideoOGL.vcproj
View File

@ -746,14 +746,6 @@
RelativePath=".\Src\VertexManager.h"
>
</File>
<File
RelativePath=".\Src\XFStructs.cpp"
>
</File>
<File
RelativePath=".\Src\XFStructs.h"
>
</File>
</Filter>
<Filter
Name="Render"

58
Source/Plugins/Plugin_VideoOGL/Src/BPStructs.cpp
View File

@ -86,7 +86,6 @@ void BPWritten(int addr, int changes, int newval)
}
else if (glIsEnabled(GL_CULL_FACE) == GL_TRUE)
glDisable(GL_CULL_FACE);
PixelShaderManager::SetGenModeChanged();
}
break;
@ -177,7 +176,6 @@ void BPWritten(int addr, int changes, int newval)
case BPMEM_BLENDMODE:
if (changes & 0xFFFF) {
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
@ -320,7 +318,7 @@ void BPWritten(int addr, int changes, int newval)
#endif
}
break;
case 0xf6: // ksel0
case 0xf7: // ksel1
case 0xf8: // ksel2
@ -343,23 +341,23 @@ void BPWritten(int addr, int changes, int newval)
{
case 0x02:
g_VideoInitialize.pSetPEFinish(); // may generate interrupt
DebugLog("GXSetDrawDone SetPEFinish (value: 0x%02X)", (newval & 0xFFFF));
DEBUG_LOG("GXSetDrawDone SetPEFinish (value: 0x%02X)", (newval & 0xFFFF));
break;
default:
DebugLog("GXSetDrawDone ??? (value 0x%02X)", (newval & 0xFFFF));
DEBUG_LOG("GXSetDrawDone ??? (value 0x%02X)", (newval & 0xFFFF));
break;
}
break;
case BPMEM_PE_TOKEN_ID:
g_VideoInitialize.pSetPEToken(static_cast<u16>(newval & 0xFFFF), FALSE);
DebugLog("SetPEToken 0x%04x", (newval & 0xFFFF));
DEBUG_LOG("SetPEToken 0x%04x", (newval & 0xFFFF));
break;
case BPMEM_PE_TOKEN_INT_ID:
g_VideoInitialize.pSetPEToken(static_cast<u16>(newval & 0xFFFF), TRUE);
DebugLog("SetPEToken + INT 0x%04x", (newval & 0xFFFF));
DEBUG_LOG("SetPEToken + INT 0x%04x", (newval & 0xFFFF));
break;
case BPMEM_SETGPMETRIC: // Set gp metric?
@ -557,6 +555,7 @@ void BPWritten(int addr, int changes, int newval)
// Not sure if it's a good idea, though. For now, we hash texture palettes
}
break;
default:
switch (addr & 0xFC) //texture sampler filter
@ -649,7 +648,7 @@ void BPWritten(int addr, int changes, int newval)
if (changes) {
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
PixelShaderManager::SetTevIndirectChanged(addr-0x10);
PixelShaderManager::SetTevIndirectChanged(addr - 0x10);
}
break;
@ -657,7 +656,7 @@ void BPWritten(int addr, int changes, int newval)
if (changes) {
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
PixelShaderManager::SetTexDimsChanged((addr>>1)&0x7);
PixelShaderManager::SetTexDimsChanged((addr >> 1) & 0x7);
}
break;
@ -667,7 +666,7 @@ void BPWritten(int addr, int changes, int newval)
{
VertexManager::Flush();
((u32*)&bpmem)[addr] = newval;
PixelShaderManager::SetTevCombinerChanged((addr&0x1f)/2);
PixelShaderManager::SetTevCombinerChanged((addr & 0x1f) / 2);
}
break;
@ -730,42 +729,3 @@ void BPWritten(int addr, int changes, int newval)
}
}
// Call browser: OpcodeDecoding.cpp ExecuteDisplayList > Decode() > LoadBPReg()
void LoadBPReg(u32 value0)
{
DVSTARTPROFILE();
// Handle the mask register
int opcode = value0 >> 24;
int oldval = ((u32*)&bpmem)[opcode];
int newval = (oldval & ~bpmem.bpMask) | (value0 & bpmem.bpMask);
// Check if it's a new value
int changes = (oldval ^ newval) & 0xFFFFFF;
//reset the mask register
if (opcode != 0xFE)
bpmem.bpMask = 0xFFFFFF;
// Notify the video handling so it can update render states
BPWritten(opcode, changes, newval);
}
// 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:
BPWritten(i, 0xFFFFFF, ((u32*)&bpmem)[i]);
}
}
}

1
Source/Plugins/Plugin_VideoOGL/Src/NativeVertexFormat.cpp
View File

@ -99,6 +99,7 @@ inline GLuint VarToGL(VarType t)
void GLVertexFormat::Initialize(const PortableVertexDeclaration &_vtx_decl)
{
vertex_stride = _vtx_decl.stride;
using namespace Gen;
if (_vtx_decl.stride & 3) {

2
Source/Plugins/Plugin_VideoOGL/Src/VertexManager.cpp
View File

@ -137,7 +137,7 @@ void Flush()
_assert_(s_pCurBufferPointer != s_pBaseBufferPointer);
#if defined(_DEBUG) || defined(DEBUGFAST)
PRIM_LOG("frame%d:\ncomps=0x%x, numchan=%d, dualtex=%d, ztex=%d, cole=%d, alpe=%d, ze=%d\n", g_Config.iSaveTargetId, xfregs.numTexGens,
PRIM_LOG("frame%d:\n texgen=%d, numchan=%d, dualtex=%d, ztex=%d, cole=%d, alpe=%d, ze=%d\n", g_Config.iSaveTargetId, xfregs.numTexGens,
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) {