1
0
mirror of https://github.com/blawar/GLideN64.git synced 2024-07-07 03:13:49 +00:00
GLideN64/RSP.cpp
2015-05-13 10:12:31 +06:00

259 lines
6.2 KiB
C++

#include <algorithm>
#include "Debug.h"
#include "RSP.h"
#include "RDP.h"
#include "N64.h"
#include "F3D.h"
#include "VI.h"
#include "Combiner.h"
#include "FrameBuffer.h"
#include "DepthBuffer.h"
#include "GBI.h"
#include "Config.h"
using namespace std;
RSPInfo RSP;
void RSP_LoadMatrix( f32 mtx[4][4], u32 address )
{
f32 recip = 1.5258789e-05f;
#ifdef WIN32_ASM
__asm {
mov esi, dword ptr [RDRAM];
add esi, dword ptr [address];
mov edi, dword ptr [mtx];
mov ecx, 4
LoadLoop:
fild word ptr [esi+02h]
movzx eax, word ptr [esi+22h]
mov dword ptr [edi], eax
fild dword ptr [edi]
fmul dword ptr [recip]
fadd
fstp dword ptr [edi]
fild word ptr [esi+00h]
movzx eax, word ptr [esi+20h]
mov dword ptr [edi+04h], eax
fild dword ptr [edi+04h]
fmul dword ptr [recip]
fadd
fstp dword ptr [edi+04h]
fild word ptr [esi+06h]
movzx eax, word ptr [esi+26h]
mov dword ptr [edi+08h], eax
fild dword ptr [edi+08h]
fmul dword ptr [recip]
fadd
fstp dword ptr [edi+08h]
fild word ptr [esi+04h]
movzx eax, word ptr [esi+24h]
mov dword ptr [edi+0Ch], eax
fild dword ptr [edi+0Ch]
fmul dword ptr [recip]
fadd
fstp dword ptr [edi+0Ch]
add esi, 08h
add edi, 10h
loop LoadLoop
}
#else // WIN32_ASM
# ifdef X86_ASM
__asm__ __volatile__(
".intel_syntax noprefix" "\n\t"
"LoadLoop:" "\n\t"
" fild word ptr [esi+0x02]" "\n\t"
" movzx eax, word ptr [esi+0x22]" "\n\t"
" mov dword ptr [edi], eax" "\n\t"
" fild dword ptr [edi]" "\n\t"
" fmul %0" "\n\t"
" fadd" "\n\t"
" fstp dword ptr [edi]" "\n\t"
" fild word ptr [esi+0x00]" "\n\t"
" movzx eax, word ptr [esi+0x20]" "\n\t"
" mov dword ptr [edi+0x04], eax" "\n\t"
" fild dword ptr [edi+0x04]" "\n\t"
" fmul %0" "\n\t"
" fadd" "\n\t"
" fstp dword ptr [edi+0x04]" "\n\t"
" fild word ptr [esi+0x06]" "\n\t"
" movzx eax, word ptr [esi+0x26]" "\n\t"
" mov dword ptr [edi+0x08], eax" "\n\t"
" fild dword ptr [edi+0x08]" "\n\t"
" fmul %0" "\n\t"
" fadd" "\n\t"
" fstp dword ptr [edi+0x08]" "\n\t"
" fild word ptr [esi+0x04]" "\n\t"
" movzx eax, word ptr [esi+0x24]" "\n\t"
" mov dword ptr [edi+0x0C], eax" "\n\t"
" fild dword ptr [edi+0x0C]" "\n\t"
" fmul %0" "\n\t"
" fadd" "\n\t"
" fstp dword ptr [edi+0x0C]" "\n\t"
" add esi, 0x08" "\n\t"
" add edi, 0x10" "\n\t"
" loop LoadLoop" "\n\t"
".att_syntax prefix" "\n\t"
: /* no output */
: "f"(recip), "S"((int)RDRAM+address), "D"(mtx), "c"(4)
: "memory" );
# else // X86_ASM
struct _N64Matrix
{
SHORT integer[4][4];
WORD fraction[4][4];
} *n64Mat = (struct _N64Matrix *)&RDRAM[address];
int i, j;
for (i = 0; i < 4; i++)
for (j = 0; j < 4; j++)
mtx[i][j] = (GLfloat)(n64Mat->integer[i][j^1]) + (GLfloat)(n64Mat->fraction[i][j^1]) * recip;
# endif // !X86_ASM
#endif // WIN32_ASM
}
void RSP_ProcessDList()
{
VI_UpdateSize();
OGL_UpdateScale();
RSP.PC[0] = *(u32*)&DMEM[0x0FF0];
RSP.PCi = 0;
RSP.count = 0;
RSP.halt = FALSE;
RSP.busy = TRUE;
gSP.matrix.stackSize = min( 32U, *(u32*)&DMEM[0x0FE4] >> 6 );
gSP.matrix.modelViewi = 0;
gSP.changed &= ~CHANGED_CPU_FB_WRITE;
gSP.changed |= CHANGED_MATRIX;
for (int i = 0; i < 4; i++)
for (int j = 0; j < 4; j++)
gSP.matrix.modelView[0][i][j] = 0.0f;
gSP.matrix.modelView[0][0][0] = 1.0f;
gSP.matrix.modelView[0][1][1] = 1.0f;
gSP.matrix.modelView[0][2][2] = 1.0f;
gSP.matrix.modelView[0][3][3] = 1.0f;
u32 uc_start = *(u32*)&DMEM[0x0FD0];
u32 uc_dstart = *(u32*)&DMEM[0x0FD8];
u32 uc_dsize = *(u32*)&DMEM[0x0FDC];
if ((uc_start != RSP.uc_start) || (uc_dstart != RSP.uc_dstart))
gSPLoadUcodeEx( uc_start, uc_dstart, uc_dsize );
gDPSetAlphaCompare( G_AC_NONE );
gDPSetDepthSource( G_ZS_PIXEL );
gDPSetRenderMode( 0, 0 );
gDPSetAlphaDither( G_AD_DISABLE );
gDPSetColorDither( G_CD_DISABLE );
gDPSetCombineKey( G_CK_NONE );
gDPSetTextureConvert( G_TC_FILT );
gDPSetTextureFilter( G_TF_POINT );
gDPSetTextureLUT( G_TT_NONE );
gDPSetTextureLOD( G_TL_TILE );
gDPSetTextureDetail( G_TD_CLAMP );
gDPSetTexturePersp( G_TP_PERSP );
gDPSetCycleType( G_CYC_1CYCLE );
gDPPipelineMode( G_PM_NPRIMITIVE );
while (!RSP.halt)
{
if ((RSP.PC[RSP.PCi] + 8) > RDRAMSize)
{
#ifdef DEBUG
switch (Debug.level)
{
case DEBUG_LOW:
DebugMsg( DEBUG_LOW | DEBUG_ERROR, "ATTEMPTING TO EXECUTE RSP COMMAND AT INVALID RDRAM LOCATION\n" );
break;
case DEBUG_MEDIUM:
DebugMsg( DEBUG_MEDIUM | DEBUG_ERROR, "Attempting to execute RSP command at invalid RDRAM location\n" );
break;
case DEBUG_HIGH:
DebugMsg( DEBUG_HIGH | DEBUG_ERROR, "// Attempting to execute RSP command at invalid RDRAM location\n" );
break;
}
#endif
break;
}
// printf( "!!!!!! RDRAM = 0x%8.8x\n", RDRAM );//RSP.PC[RSP.PCi] );
/* {
static u8 *lastRDRAM = 0;
if (lastRDRAM == 0)
lastRDRAM = RDRAM;
if (RDRAM != lastRDRAM)
{
__asm__( "int $3" );
}
}*/
u32 w0 = *(u32*)&RDRAM[RSP.PC[RSP.PCi]];
u32 w1 = *(u32*)&RDRAM[RSP.PC[RSP.PCi] + 4];
RSP.cmd = _SHIFTR( w0, 24, 8 );
#ifdef DEBUG
DebugRSPState( RSP.PCi, RSP.PC[RSP.PCi], _SHIFTR( w0, 24, 8 ), w0, w1 );
DebugMsg( DEBUG_LOW | DEBUG_HANDLED, "0x%08lX: CMD=0x%02lX W0=0x%08lX W1=0x%08lX\n", RSP.PC[RSP.PCi], _SHIFTR( w0, 24, 8 ), w0, w1 );
#endif
RSP.PC[RSP.PCi] += 8;
RSP.nextCmd = _SHIFTR( *(u32*)&RDRAM[RSP.PC[RSP.PCi]], 24, 8 );
GBI.cmd[RSP.cmd]( w0, w1 );
}
if (config.frameBufferEmulation.copyToRDRAM)
FrameBuffer_CopyToRDRAM( gDP.colorImage.address, false );
if (config.frameBufferEmulation.copyDepthToRDRAM)
FrameBuffer_CopyDepthBuffer( gDP.colorImage.address );
RSP.busy = FALSE;
RSP.DList++;
gSP.changed |= CHANGED_COLORBUFFER;
}
void RSP_Init()
{
#ifdef _WINDOWS
// Calculate RDRAM size by intentionally causing an access violation
u32 test;
try
{
test = RDRAM[0x007FFFFF] + 1;
}
catch (...)
{
test = 0;
}
if (test > 0)
RDRAMSize = 0x7FFFFF;
else
RDRAMSize = 0x3FFFFF;
#else // _WINDOWS
RDRAMSize = 1024 * 1024 * 8;
#endif // _WINDOWS
RSP.DList = 0;
RSP.uc_start = RSP.uc_dstart = 0;
gDP.loadTile = &gDP.tiles[7];
gSP.textureTile[0] = &gDP.tiles[0];
gSP.textureTile[1] = &gDP.tiles[1];
// DepthBuffer_Init();
GBI_Init();
OGL_Start();
}