mixed commit:

in D3D and Opengl:
fixed one nasty bug in texture loading where if a dynamic texture keeps his format but the tlut format is changed, the try or reloading the texture in the same texture could cause a hang if the size of the resulting texture is different than the original (size in bytes)
Applied a ugly temporal hack to the texture conversor to solve efb to ram misalignments and effect distortions.
in D3D:
Pseudo implementation of logic ops using basic blending: the first 8 operations are "good approximations", the remaining 8 are bullshit :) if someone have a better approximation to emulate this logic please let me know.
please test if i don't break anything in the process and test Mario kart wee you will get a nice surprise.:)
 

git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4656 8ced0084-cf51-0410-be5f-012b33b47a6e
This commit is contained in:
Rodolfo Osvaldo Bogado 2009-12-07 18:48:31 +00:00
parent 85a9b3dc2c
commit d02426a8e9
14 changed files with 634 additions and 67 deletions

View File

@ -545,7 +545,7 @@ const char *GeneratePixelShader(u32 texture_mask, bool dstAlphaEnable, u32 HLSL)
// alpha test will always fail, so restart the shader and just make it an empty function
p = pmainstart;
WRITE(p, "ocol0 = 0;\n");
WRITE(p, HLSL ? "clip(-1);" : "discard;\n");
WRITE(p, "discard;return;\n");
}
else
{
@ -582,9 +582,9 @@ static const char *TEVCMPColorOPTable[16] =
"float3(0.0f,0.0f,0.0f)",//7
" %s + ((%s.r > %s.r + (1.0f/510.0f)) ? %s : float3(0.0f,0.0f,0.0f))",//#define TEVCMP_R8_GT 8
" %s + ((abs(%s.r - %s.r) < (1.0f/255.0f)) ? %s : float3(0.0f,0.0f,0.0f))",//#define TEVCMP_R8_EQ 9
" %s + (( dot(%s.rgb, comp16) >= (dot(%s.rgb, comp16) + (1.0f/510.0f))) ? %s : float3(0.0f,0.0f,0.0f))",//#define TEVCMP_GR16_GT 10
" %s + (( dot(%s.rgb, comp16) > (dot(%s.rgb, comp16) + (1.0f/510.0f))) ? %s : float3(0.0f,0.0f,0.0f))",//#define TEVCMP_GR16_GT 10
" %s + (abs(dot(%s.rgb, comp16) - dot(%s.rgb, comp16)) < (1.0f/255.0f) ? %s : float3(0.0f,0.0f,0.0f))",//#define TEVCMP_GR16_EQ 11
" %s + (( dot(%s.rgb, comp24) >= (dot(%s.rgb, comp24) + (1.0f/510.0f))) ? %s : float3(0.0f,0.0f,0.0f))",//#define TEVCMP_BGR24_GT 12
" %s + (( dot(%s.rgb, comp24) > (dot(%s.rgb, comp24) + (1.0f/510.0f))) ? %s : float3(0.0f,0.0f,0.0f))",//#define TEVCMP_BGR24_GT 12
" %s + (abs(dot(%s.rgb, comp24) - dot(%s.rgb, comp24)) < (1.0f/255.0f) ? %s : float3(0.0f,0.0f,0.0f))",//#define TEVCMP_BGR24_EQ 13
" %s + (max(sign(%s.rgb - %s.rgb - (1.0f/510.0f)),float3(0.0f,0.0f,0.0f)) * %s)",//#define TEVCMP_RGB8_GT 14
" %s + ((float3(1.0f,1.0f,1.0f) - max(sign(abs(%s.rgb - %s.rgb) - (1.0f/255.0f)),float3(0.0f,0.0f,0.0f))) * %s)"//#define TEVCMP_RGB8_EQ 15
@ -601,13 +601,13 @@ static const char *TEVCMPAlphaOPTable[16] =
"0.0f",//5
"0.0f",//6
"0.0f",//7
" %s + ((%s.r >= (%s.r + (1.0f/510.0f))) ? %s : 0.0f)",//#define TEVCMP_R8_GT 8
" %s + ((%s.r > (%s.r + (1.0f/510.0f))) ? %s : 0.0f)",//#define TEVCMP_R8_GT 8
" %s + (abs(%s.r - %s.r) < (1.0f/255.0f) ? %s : 0.0f)",//#define TEVCMP_R8_EQ 9
" %s + ((dot(%s.rgb, comp16) >= (dot(%s.rgb, comp16) + (1.0f/510.0f))) ? %s : 0.0f)",//#define TEVCMP_GR16_GT 10
" %s + ((dot(%s.rgb, comp16) > (dot(%s.rgb, comp16) + (1.0f/510.0f))) ? %s : 0.0f)",//#define TEVCMP_GR16_GT 10
" %s + (abs(dot(%s.rgb, comp16) - dot(%s.rgb, comp16)) < (1.0f/255.0f) ? %s : 0.0f)",//#define TEVCMP_GR16_EQ 11
" %s + ((dot(%s.rgb, comp24) >= (dot(%s.rgb, comp24) + (1.0f/510.0f))) ? %s : 0.0f)",//#define TEVCMP_BGR24_GT 12
" %s + ((dot(%s.rgb, comp24) > (dot(%s.rgb, comp24) + (1.0f/510.0f))) ? %s : 0.0f)",//#define TEVCMP_BGR24_GT 12
" %s + (abs(dot(%s.rgb, comp24) - dot(%s.rgb, comp24)) < (1.0f/255.0f) ? %s : 0.0f)",//#define TEVCMP_BGR24_EQ 13
" %s + ((%s.a >= (%s.a + (1.0f/510.0f))) ? %s : 0.0f)",//#define TEVCMP_A8_GT 14
" %s + ((%s.a > (%s.a + (1.0f/510.0f))) ? %s : 0.0f)",//#define TEVCMP_A8_GT 14
" %s + (abs(%s.a - %s.a) < (1.0f/255.0f) ? %s : 0.0f)"//#define TEVCMP_A8_EQ 15
};
@ -862,6 +862,8 @@ static const char *tevAlphaFuncsTable[] =
"(true)" //ALPHACMP_ALWAYS 7
};
static const char *tevAlphaFunclogicTable[] =
{
" && ", // and
@ -903,10 +905,7 @@ static bool WriteAlphaTest(char *&p, u32 HLSL)
// Seems we need discard for Cg and clip for d3d. sigh.
if (HLSL)
WRITE(p, "clip( (");
else
WRITE(p, "discard(!( ");
WRITE(p, "if(!( ");
int compindex = bpmem.alphaFunc.comp0 % 8;
WRITE(p, tevAlphaFuncsTable[compindex],alphaRef[0]);//lookup the first component from the alpha function table
@ -916,12 +915,8 @@ static bool WriteAlphaTest(char *&p, u32 HLSL)
compindex = bpmem.alphaFunc.comp1 % 8;
WRITE(p, tevAlphaFuncsTable[compindex],alphaRef[1]);//lookup the second component from the alpha function table
if (HLSL) {
// clip works differently than discard - discard takes a bool, clip takes a value that kills the pixel on negative
WRITE(p, ") ? 1 : -1);\n");
} else {
WRITE(p, "));\n");
}
WRITE(p, ")){ocol0 = 0;discard;return;}\n");
return true;
}
@ -967,6 +962,6 @@ static void WriteFog(char *&p)
WARN_LOG(VIDEO, "Unknown Fog Type! %08x", bpmem.fog.c_proj_fsel.fsel);
}
WRITE(p, " prev.rgb = (1.0f - fog) * prev.rgb + (fog * "I_FOG"[0].rgb);\n");
WRITE(p, " prev.rgb = lerp(prev.rgb,"I_FOG"[0].rgb,fog);\n");
}

View File

@ -107,12 +107,16 @@ void WriteSwizzler(char*& p, u32 format,bool HLSL)
if(!HLSL)
WRITE(p," sampleUv.y = textureDims.y - sampleUv.y;\n");
WRITE(p, " sampleUv.x = sampleUv.x + textureDims.z;\n"
" sampleUv.y = sampleUv.y + textureDims.w;\n");
WRITE(p, " sampleUv = sampleUv + textureDims.zw;\n");
if(HLSL)
{
WRITE(p, " sampleUv.x = sampleUv.x / blkDims.z;\n"
" sampleUv.y = sampleUv.y / blkDims.w;\n");
WRITE(p, " sampleUv = sampleUv + float2(1.0f,1.0f);\n"
" sampleUv = sampleUv / blkDims.zw;\n");
}
else
{
WRITE(p, " sampleUv = sampleUv + float2(1.0f,-1.0f);\n");
}
}
@ -158,15 +162,18 @@ void Write32BitSwizzler(char*& p, u32 format, bool HLSL)
WRITE(p, " sampleUv = sampleUv * blkDims.xy;\n");
if(!HLSL)
WRITE(p, " sampleUv.y = textureDims.y - sampleUv.y;\n");
WRITE(p," sampleUv.y = textureDims.y - sampleUv.y;\n");
WRITE(p, " sampleUv.x = sampleUv.x + textureDims.z;\n");
WRITE(p, " sampleUv.y = sampleUv.y + textureDims.w;\n");
WRITE(p, " sampleUv = sampleUv + textureDims.zw;\n");
if(HLSL)
{
WRITE(p, " sampleUv.x = sampleUv.x / blkDims.z;\n"
" sampleUv.y = sampleUv.y / blkDims.w;\n");
WRITE(p, " sampleUv = sampleUv + float2(1.0f,1.0f);\n"
" sampleUv = sampleUv / blkDims.zw;\n");
}
else
{
WRITE(p, " sampleUv = sampleUv + float2(1.0f,-1.0f);\n");
}
}

View File

@ -418,6 +418,219 @@ static PC_TexFormat GetPCFormatFromTLUTFormat(int tlutfmt)
return PC_TEX_FMT_NONE; // Error
}
PC_TexFormat GetPC_TexFormat(int texformat, int tlutfmt)
{
switch (texformat)
{
case GX_TF_C4:
return GetPCFormatFromTLUTFormat(tlutfmt);
case GX_TF_I4:
return PC_TEX_FMT_I4_AS_I8;
case GX_TF_I8: // speed critical
return PC_TEX_FMT_I8;
case GX_TF_C8:
return GetPCFormatFromTLUTFormat(tlutfmt);
case GX_TF_IA4:
return PC_TEX_FMT_IA4_AS_IA8;
case GX_TF_IA8:
return PC_TEX_FMT_IA8;
case GX_TF_C14X2:
return GetPCFormatFromTLUTFormat(tlutfmt);
case GX_TF_RGB565:
return PC_TEX_FMT_RGB565;
case GX_TF_RGB5A3:
return PC_TEX_FMT_BGRA32;
case GX_TF_RGBA8: // speed critical
return PC_TEX_FMT_BGRA32;
case GX_TF_CMPR: // speed critical
// The metroid games use this format almost exclusively.
{
return PC_TEX_FMT_BGRA32;
}
}
// The "copy" texture formats, too?
return PC_TEX_FMT_NONE;
}
PC_TexFormat TexDecoder_DirectDecode_real(u8 *dst, const u8 *src, int width, int height,int Pitch, int texformat, int tlutaddr, int tlutfmt)
{
switch (texformat)
{
case GX_TF_C4:
if (tlutfmt == 2)
{
// Special decoding is required for TLUT format 5A3
for (int y = 0; y < height; y += 8)
for (int x = 0; x < width; x += 8)
for (int iy = 0; iy < 8; iy++, src += 4)
decodebytesC4_5A3_To_BGRA32((u32*)dst + (y + iy) * width + x, src, tlutaddr);
}
else
{
for (int y = 0; y < height; y += 8)
for (int x = 0; x < width; x += 8)
for (int iy = 0; iy < 8; iy++, src += 4)
decodebytesC4_To_Raw16((u16*)dst + (y + iy) * width + x, src, tlutaddr);
}
return GetPCFormatFromTLUTFormat(tlutfmt);
case GX_TF_I4:
{
for (int y = 0; y < height; y += 8)
for (int x = 0; x < width; x += 8)
for (int iy = 0; iy < 8; iy++, src += 4)
for (int ix = 0; ix < 4; ix++)
{
int val = src[ix];
dst[(y + iy) * width + x + ix * 2] = Convert4To8(val >> 4);
dst[(y + iy) * width + x + ix * 2 + 1] = Convert4To8(val & 0xF);
}
}
return PC_TEX_FMT_I4_AS_I8;
case GX_TF_I8: // speed critical
{
for (int y = 0; y < height; y += 4)
for (int x = 0; x < width; x += 8)
for (int iy = 0; iy < 4; iy++, src += 8)
memcpy(dst + (y + iy)*width+x, src, 8);
}
return PC_TEX_FMT_I8;
case GX_TF_C8:
if (tlutfmt == 2)
{
// Special decoding is required for TLUT format 5A3
for (int y = 0; y < height; y += 4)
for (int x = 0; x < width; x += 8)
for (int iy = 0; iy < 4; iy++, src += 8)
decodebytesC8_5A3_To_BGRA32((u32*)dst + (y + iy) * width + x, src, tlutaddr);
}
else
{
for (int y = 0; y < height; y += 4)
for (int x = 0; x < width; x += 8)
for (int iy = 0; iy < 4; iy++, src += 8)
decodebytesC8_To_Raw16((u16*)dst + (y + iy) * width + x, src, tlutaddr);
}
return GetPCFormatFromTLUTFormat(tlutfmt);
case GX_TF_IA4:
{
for (int y = 0; y < height; y += 4)
for (int x = 0; x < width; x += 8)
for (int iy = 0; iy < 4; iy++, src += 8)
decodebytesIA4((u16*)dst + (y + iy) * width + x, src);
}
return PC_TEX_FMT_IA4_AS_IA8;
case GX_TF_IA8:
{
for (int y = 0; y < height; y += 4)
for (int x = 0; x < width; x += 4)
for (int iy = 0; iy < 4; iy++, src += 8)
{
u16 *ptr = (u16 *)dst + (y + iy) * width + x;
u16 *s = (u16 *)src;
for(int j = 0; j < 4; j++)
*ptr++ = Common::swap16(*s++);
}
}
return PC_TEX_FMT_IA8;
case GX_TF_C14X2:
if (tlutfmt == 2)
{
// Special decoding is required for TLUT format 5A3
for (int y = 0; y < height; y += 4)
for (int x = 0; x < width; x += 4)
for (int iy = 0; iy < 4; iy++, src += 8)
decodebytesC14X2_5A3_To_BGRA32((u32*)dst + (y + iy) * width + x, (u16*)src, tlutaddr);
}
else
{
for (int y = 0; y < height; y += 4)
for (int x = 0; x < width; x += 4)
for (int iy = 0; iy < 4; iy++, src += 8)
decodebytesC14X2_To_Raw16((u16*)dst + (y + iy) * width + x, (u16*)src, tlutaddr);
}
return GetPCFormatFromTLUTFormat(tlutfmt);
case GX_TF_RGB565:
{
for (int y = 0; y < height; y += 4)
for (int x = 0; x < width; x += 4)
for (int iy = 0; iy < 4; iy++, src += 8)
{
u16 *ptr = (u16 *)dst + (y + iy) * width + x;
u16 *s = (u16 *)src;
for(int j = 0; j < 4; j++)
*ptr++ = Common::swap16(*s++);
}
}
return PC_TEX_FMT_RGB565;
case GX_TF_RGB5A3:
{
for (int y = 0; y < height; y += 4)
for (int x = 0; x < width; x += 4)
for (int iy = 0; iy < 4; iy++, src += 8)
//decodebytesRGB5A3((u32*)dst+(y+iy)*width+x, (u16*)src, 4);
decodebytesRGB5A3((u32*)dst+(y+iy)*width+x, (u16*)src);
}
return PC_TEX_FMT_BGRA32;
case GX_TF_RGBA8: // speed critical
{
for (int y = 0; y < height; y += 4)
for (int x = 0; x < width; x += 4)
{
for (int iy = 0; iy < 4; iy++)
decodebytesARGB8_4((u32*)dst + (y+iy)*width + x, (u16*)src + 4 * iy, (u16*)src + 4 * iy + 16);
src += 64;
}
}
return PC_TEX_FMT_BGRA32;
case GX_TF_CMPR: // speed critical
// The metroid games use this format almost exclusively.
{
#if 0 // TODO - currently does not handle transparency correctly and causes problems when texture dimensions are not multiples of 8
// 11111111 22222222 55555555 66666666
// 33333333 44444444 77777777 88888888
for (int y = 0; y < height; y += 8)
{
for (int x = 0; x < width; x += 8)
{
copyDXTBlock(dst+(y/2)*width+x*2, src);
src += 8;
copyDXTBlock(dst+(y/2)*width+x*2+8, src);
src += 8;
copyDXTBlock(dst+(y/2+2)*width+x*2, src);
src += 8;
copyDXTBlock(dst+(y/2+2)*width+x*2+8, src);
src += 8;
}
}
return PC_TEX_FMT_DXT1;
#else
for (int y = 0; y < height; y += 8)
{
for (int x = 0; x < width; x += 8)
{
decodeDXTBlock((u32*)dst + y * width + x, (DXTBlock*)src, width);
src += sizeof(DXTBlock);
decodeDXTBlock((u32*)dst + y * width + x + 4, (DXTBlock*)src, width);
src += sizeof(DXTBlock);
decodeDXTBlock((u32*)dst + (y + 4) * width + x, (DXTBlock*)src, width);
src += sizeof(DXTBlock);
decodeDXTBlock((u32*)dst + (y + 4) * width + x + 4, (DXTBlock*)src, width);
src += sizeof(DXTBlock);
}
}
#endif
return PC_TEX_FMT_BGRA32;
}
}
// The "copy" texture formats, too?
return PC_TEX_FMT_NONE;
}
//switch endianness, unswizzle
//TODO: to save memory, don't blindly convert everything to argb8888
@ -599,12 +812,91 @@ PC_TexFormat TexDecoder_Decode_real(u8 *dst, const u8 *src, int width, int heigh
return PC_TEX_FMT_NONE;
}
void TexDecoder_SetTexFmtOverlayOptions(bool enable, bool center)
{
TexFmt_Overlay_Enable = enable;
TexFmt_Overlay_Center = center;
}
void TexDecoder_DirectDecode(u8 *dst, const u8 *src, int width, int height,int Pitch, int texformat, int tlutaddr, int tlutfmt)
{
PC_TexFormat retval = TexDecoder_DirectDecode_real(dst,src,width,height,Pitch,texformat,tlutaddr,tlutfmt);
if ((!TexFmt_Overlay_Enable)|| (retval == PC_TEX_FMT_NONE))
return;
int w = min(width, 40);
int h = min(height, 10);
int xoff = (width - w) >> 1;
int yoff = (height - h) >> 1;
if (!TexFmt_Overlay_Center)
{
xoff=0;
yoff=0;
}
const char* fmt = texfmt[texformat&15];
while (*fmt)
{
int xcnt = 0;
int nchar = sfont_map[(int)*fmt];
const unsigned char *ptr = sfont_raw[nchar]; // each char is up to 9x10
for (int x = 0; x < 9;x++)
{
if (ptr[x] == 0x78)
break;
xcnt++;
}
for (int y=0; y < 10; y++)
{
for (int x=0; x < xcnt; x++)
{
switch(retval)
{
case PC_TEX_FMT_I8:
{
// TODO: Is this an acceptable way to draw in I8?
u8 *dtp = (u8*)dst;
dtp[(y + yoff) * width + x + xoff] = ptr[x] ? 0xFF : 0x88;
break;
}
case PC_TEX_FMT_IA8:
case PC_TEX_FMT_IA4_AS_IA8:
{
u16 *dtp = (u16*)dst;
dtp[(y + yoff) * width + x + xoff] = ptr[x] ? 0xFFFF : 0xFF00;
break;
}
case PC_TEX_FMT_RGB565:
{
u16 *dtp = (u16*)dst;
dtp[(y + yoff)*width + x + xoff] = ptr[x] ? 0xFFFF : 0x0000;
break;
}
default:
case PC_TEX_FMT_BGRA32:
{
int *dtp = (int*)dst;
dtp[(y + yoff) * width + x + xoff] = ptr[x] ? 0xFFFFFFFF : 0xFF000000;
break;
}
}
}
ptr += 9;
}
xoff += xcnt;
fmt++;
}
}
PC_TexFormat TexDecoder_Decode(u8 *dst, const u8 *src, int width, int height, int texformat, int tlutaddr, int tlutfmt)
{
#if defined(HAVE_OPENCL) && HAVE_OPENCL

View File

@ -84,9 +84,11 @@ enum PC_TexFormat
};
PC_TexFormat TexDecoder_Decode(u8 *dst, const u8 *src, int width, int height, int texformat, int tlutaddr, int tlutfmt);
void TexDecoder_DirectDecode(u8 *dst, const u8 *src, int width, int height,int Pitch, int texformat, int tlutaddr, int tlutfmt);
PC_TexFormat GetPC_TexFormat(int texformat, int tlutfmt);
void TexDecoder_DecodeTexel(u8 *dst, const u8 *src, int s, int t, int imageWidth, int texformat, int tlutaddr, int tlutfmt);
u32 TexDecoder_GetSafeTextureHash(const u8 *src, int width, int height, int texformat, u32 seed=0);
u32 TexDecoder_GetTlutHash(const u8* src, int len);

View File

@ -71,7 +71,7 @@ void SetDitherMode(const BPCmd &bp)
}
void SetLogicOpMode(const BPCmd &bp)
{
// Logic op blending. D3D can't do this but can fake some modes.
Renderer::SetLogicOpMode();
}
void SetColorMask(const BPCmd &bp)

View File

@ -103,13 +103,20 @@ LPDIRECT3DTEXTURE9 CreateTexture2D(const u8* buffer, const int width, const int
break;
case D3DFMT_A8R8G8B8:
{
u32* pIn = pBuffer;
for (int y = 0; y < height; y++)
/*if(Lock.Pitch == width * 4)
{
u32* pBits = (u32*)((u8*)Lock.pBits + (y * Lock.Pitch));
memcpy(pBits, pIn, width * 4);
pIn += pitch;
memcpy(Lock.pBits,buffer,width*height*4);
}
else
{*/
u32* pIn = pBuffer;
for (int y = 0; y < height; y++)
{
u32* pBits = (u32*)((u8*)Lock.pBits + (y * Lock.Pitch));
memcpy(pBits, pIn, width * 4);
pIn += pitch;
}
//}
}
break;
case D3DFMT_DXT1:
@ -122,6 +129,25 @@ LPDIRECT3DTEXTURE9 CreateTexture2D(const u8* buffer, const int width, const int
return pTexture;
}
LPDIRECT3DTEXTURE9 CreateOnlyTexture2D(const int width, const int height, D3DFORMAT fmt)
{
LPDIRECT3DTEXTURE9 pTexture;
// crazy bitmagic, sorry :)
bool isPow2 = !((width&(width-1)) || (height&(height-1)));
bool bExpand = false;
HRESULT hr;
// TODO(ector): Allow mipmaps for non-pow textures on newer cards?
// TODO(ector): Use the game-specified mipmaps?
if (!isPow2)
hr = dev->CreateTexture(width, height, 1, 0, fmt, D3DPOOL_MANAGED, &pTexture, NULL);
else
hr = dev->CreateTexture(width, height, 0, D3DUSAGE_AUTOGENMIPMAP, fmt, D3DPOOL_MANAGED, &pTexture, NULL);
if (FAILED(hr))
return 0;
return pTexture;
}
void ReplaceTexture2D(LPDIRECT3DTEXTURE9 pTexture, const u8* buffer, const int width, const int height,const int pitch, D3DFORMAT fmt)
{
u32* pBuffer = (u32*)buffer;

View File

@ -25,4 +25,5 @@ namespace D3D
void ReplaceTexture2D(LPDIRECT3DTEXTURE9 pTexture, const u8* buffer, const int width, const int height,const int pitch, D3DFORMAT fmt);
LPDIRECT3DTEXTURE9 CreateRenderTarget(const int width, const int height);
LPDIRECT3DSURFACE9 CreateDepthStencilSurface(const int width, const int height);
LPDIRECT3DTEXTURE9 CreateOnlyTexture2D(const int width, const int height, D3DFORMAT fmt);
}

View File

@ -91,6 +91,86 @@ static const D3DBLEND d3dDestFactors[8] =
D3DBLEND_INVDESTALPHA
};
// 0 0x00
// 1 Source & destination
// 2 Source & ~destination
// 3 Source
// 4 ~Source & destination
// 5 Destination
// 6 Source ^ destination = Source & ~destination | ~Source & destination
// 7 Source | destination
// 8 ~(Source | destination)
// 9 ~(Source ^ destination) = ~Source & ~destination | Source & destination
// 10 ~Destination
// 11 Source | ~destination
// 12 ~Source
// 13 ~Source | destination
// 14 ~(Source & destination)
// 15 0xff
static const D3DBLENDOP d3dLogincOPop[16] =
{
D3DBLENDOP_ADD,
D3DBLENDOP_ADD,
D3DBLENDOP_ADD,
D3DBLENDOP_ADD,
D3DBLENDOP_ADD,
D3DBLENDOP_ADD,
D3DBLENDOP_ADD,
D3DBLENDOP_ADD,
D3DBLENDOP_REVSUBTRACT,
D3DBLENDOP_REVSUBTRACT,
D3DBLENDOP_SUBTRACT,
D3DBLENDOP_SUBTRACT,
D3DBLENDOP_REVSUBTRACT,
D3DBLENDOP_REVSUBTRACT,
D3DBLENDOP_SUBTRACT,
D3DBLENDOP_ADD
};
static const D3DBLEND d3dLogicOpSrcFactors[16] =
{
D3DBLEND_ZERO,
D3DBLEND_DESTCOLOR,
D3DBLEND_INVDESTCOLOR,
D3DBLEND_ONE,
D3DBLEND_ZERO,
D3DBLEND_ZERO,
D3DBLEND_INVDESTCOLOR,
D3DBLEND_ONE,
D3DBLEND_ONE,
D3DBLEND_INVDESTCOLOR,
D3DBLEND_ONE,
D3DBLEND_ONE,
D3DBLEND_ONE,
D3DBLEND_ONE,
D3DBLEND_ONE,
D3DBLEND_ONE
};
static const D3DBLEND d3dLogicOpDestFactors[16] =
{
D3DBLEND_ZERO,
D3DBLEND_ZERO,
D3DBLEND_ZERO,
D3DBLEND_ZERO,
D3DBLEND_INVSRCCOLOR,
D3DBLEND_ONE,
D3DBLEND_INVSRCCOLOR,
D3DBLEND_ONE,
D3DBLEND_ONE,
D3DBLEND_SRCCOLOR,
D3DBLEND_ONE,
D3DBLEND_ONE,
D3DBLEND_ONE,
D3DBLEND_ONE,
D3DBLEND_SRCCOLOR,
D3DBLEND_ONE
};
static const D3DCULL d3dCullModes[4] =
{
@ -823,10 +903,10 @@ void Renderer::ClearScreen(const EFBRectangle& rc, bool colorEnable, bool alphaE
D3D::dev->SetScissorRect(&sirc);
if(zEnable)
D3D::SetRenderState(D3DRS_ZFUNC, D3DCMP_ALWAYS);
D3D::drawClearQuad(&sirc,color,(z & 0xFFFFFF) / float(0xFFFFFF),PixelShaderCache::GetClearProgram(),VertexShaderCache::GetSimpleVertexShader());
D3D::drawClearQuad(&sirc,color ,(z & 0xFFFFFF) / float(0xFFFFFF),PixelShaderCache::GetClearProgram(),VertexShaderCache::GetSimpleVertexShader());
if(zEnable)
D3D::SetRenderState(D3DRS_ZFUNC, d3dCmpFuncs[bpmem.zmode.func]);
//D3D::dev->Clear(0, NULL, (colorEnable ? D3DCLEAR_TARGET : 0)| ( zEnable ? D3DCLEAR_ZBUFFER : 0), color,(z & 0xFFFFFF) / float(0xFFFFFF), 0);
//D3D::dev->Clear(0, NULL, (colorEnable ? D3DCLEAR_TARGET : 0)| ( zEnable ? D3DCLEAR_ZBUFFER : 0), color | ((alphaEnable)?0:0xFF000000),(z & 0xFFFFFF) / float(0xFFFFFF), 0);
SetScissorRect();
UpdateViewport();
}
@ -838,7 +918,8 @@ void Renderer::SetBlendMode(bool forceUpdate)
// 2 - reverse subtract enable (else add)
// 3-5 - srcRGB function
// 6-8 - dstRGB function
if(bpmem.blendmode.logicopenable)
return;
u32 newval = bpmem.blendmode.subtract << 2;
if (bpmem.blendmode.subtract) {
@ -937,27 +1018,12 @@ void Renderer::RestoreAPIState()
D3D::SetRenderState(D3DRS_SCISSORTESTENABLE, TRUE);
SetScissorRect();
SetColorMask();
SetBlendMode(true);
SetLogicOpMode();
}
void Renderer::SetGenerationMode()
{
D3D::SetRenderState(D3DRS_CULLMODE, d3dCullModes[bpmem.genMode.cullmode]);
if (bpmem.genMode.cullmode == 3)
{
D3D::SetRenderState(D3DRS_COLORWRITEENABLE, 0);
}
else
{
DWORD write = 0;
if (bpmem.blendmode.alphaupdate)
write = D3DCOLORWRITEENABLE_ALPHA;
if (bpmem.blendmode.colorupdate)
write |= D3DCOLORWRITEENABLE_RED | D3DCOLORWRITEENABLE_GREEN | D3DCOLORWRITEENABLE_BLUE;
D3D::SetRenderState(D3DRS_COLORWRITEENABLE, write);
}
}
void Renderer::SetDepthMode()
@ -977,7 +1043,17 @@ void Renderer::SetDepthMode()
void Renderer::SetLogicOpMode()
{
//TODO
if (bpmem.blendmode.logicopenable)
{
D3D::SetRenderState(D3DRS_ALPHABLENDENABLE, 1);
D3D::SetRenderState(D3DRS_BLENDOP, d3dLogincOPop[bpmem.blendmode.logicmode]);
D3D::SetRenderState(D3DRS_SRCBLEND, d3dLogicOpSrcFactors[bpmem.blendmode.logicmode]);
D3D::SetRenderState(D3DRS_DESTBLEND, d3dLogicOpDestFactors[bpmem.blendmode.logicmode]);
}
else
{
SetBlendMode(true);
}
}
void Renderer::SetDitherMode()

View File

@ -133,7 +133,7 @@ void TextureCache::Cleanup()
}
}
TextureCache::TCacheEntry *TextureCache::Load(int stage, u32 address, int width, int height, int tex_format, int tlutaddr, int tlutfmt)
/*TextureCache::TCacheEntry *TextureCache::Load(int stage, u32 address, int width, int height, int tex_format, int tlutaddr, int tlutfmt)
{
if (address == 0)
return NULL;
@ -203,6 +203,162 @@ TextureCache::TCacheEntry *TextureCache::Load(int stage, u32 address, int width,
}
}
//PC_TexFormat pcfmt = TexDecoder_Decode(temp, ptr, expandedWidth, height, tex_format, tlutaddr, tlutfmt);
PC_TexFormat pcfmt = GetPC_TexFormat(tex_format, tlutfmt);
D3DFORMAT d3d_fmt;
switch (pcfmt) {
case PC_TEX_FMT_BGRA32:
case PC_TEX_FMT_RGBA32:
d3d_fmt = D3DFMT_A8R8G8B8;
break;
case PC_TEX_FMT_RGB565:
d3d_fmt = D3DFMT_R5G6B5;
break;
case PC_TEX_FMT_IA4_AS_IA8:
d3d_fmt = D3DFMT_A8L8;
break;
case PC_TEX_FMT_I8:
case PC_TEX_FMT_I4_AS_I8:
d3d_fmt = D3DFMT_A8L8;
break;
case PC_TEX_FMT_IA8:
d3d_fmt = D3DFMT_A8L8;
break;
case PC_TEX_FMT_DXT1:
d3d_fmt = D3DFMT_DXT1;
break;
}
//Make an entry in the table
TCacheEntry& entry = textures[texID];
entry.oldpixel = ((u32 *)ptr)[0];
if (g_ActiveConfig.bSafeTextureCache)
entry.hash = hash_value;
else
{
entry.hash = (u32)(((double)rand() / RAND_MAX) * 0xFFFFFFFF);
((u32 *)ptr)[0] = entry.hash;
}
entry.addr = address;
entry.size_in_bytes = TexDecoder_GetTextureSizeInBytes(width, height, tex_format);
entry.isRenderTarget = false;
entry.isNonPow2 = ((width & (width - 1)) || (height & (height - 1)));
if (!skip_texture_create) {
entry.texture = D3D::CreateOnlyTexture2D(width, height, d3d_fmt);
}
D3DLOCKED_RECT Lock;
entry.texture->LockRect(0, &Lock, NULL, 0);
TexDecoder_DirectDecode((u8*)Lock.pBits,ptr,expandedWidth,height,Lock.Pitch,tex_format,tlutaddr,tlutfmt);
entry.texture->UnlockRect(0);
entry.frameCount = frameCount;
entry.w = width;
entry.h = height;
entry.fmt = tex_format;
if (g_ActiveConfig.bDumpTextures)
{
// dump texture to file
char szTemp[MAX_PATH];
char szDir[MAX_PATH];
const char* uniqueId = globals->unique_id;
bool bCheckedDumpDir = false;
sprintf(szDir, "%s/%s", FULL_DUMP_TEXTURES_DIR, uniqueId);
if (!bCheckedDumpDir)
{
if (!File::Exists(szDir) || !File::IsDirectory(szDir))
File::CreateDir(szDir);
bCheckedDumpDir = true;
}
sprintf(szTemp, "%s/%s_%08x_%i.png", szDir, uniqueId, texHash, tex_format);
//sprintf(szTemp, "%s\\txt_%04i_%i.png", g_Config.texDumpPath.c_str(), counter++, format); <-- Old method
if (!File::Exists(szTemp))
D3DXSaveTextureToFileA(szTemp,D3DXIFF_BMP,entry.texture,0);
}
INCSTAT(stats.numTexturesCreated);
SETSTAT(stats.numTexturesAlive, (int)textures.size());
//Set the texture!
D3D::SetTexture(stage, entry.texture);
DEBUGGER_PAUSE_LOG_AT(NEXT_NEW_TEXTURE,true,{printf("A new texture (%d x %d) is loaded", width, height);});
return &entry;
}*/
TextureCache::TCacheEntry *TextureCache::Load(int stage, u32 address, int width, int height, int tex_format, int tlutaddr, int tlutfmt)
{
if (address == 0)
return NULL;
u8 *ptr = g_VideoInitialize.pGetMemoryPointer(address);
int bsw = TexDecoder_GetBlockWidthInTexels(tex_format) - 1; //TexelSizeInNibbles(format)*width*height/16;
int bsh = TexDecoder_GetBlockHeightInTexels(tex_format) - 1; //TexelSizeInNibbles(format)*width*height/16;
int expandedWidth = (width + bsw) & (~bsw);
int expandedHeight = (height + bsh) & (~bsh);
u32 hash_value;
u32 texID = address;
u32 texHash;
if (g_ActiveConfig.bSafeTextureCache || g_ActiveConfig.bDumpTextures)
{
texHash = TexDecoder_GetSafeTextureHash(ptr, expandedWidth, expandedHeight, tex_format, 0);
if (g_ActiveConfig.bSafeTextureCache)
hash_value = texHash;
if ((tex_format == GX_TF_C4) || (tex_format == GX_TF_C8) || (tex_format == GX_TF_C14X2))
{
// WARNING! texID != address now => may break CopyRenderTargetToTexture (cf. TODO up)
// tlut size can be up to 32768B (GX_TF_C14X2) but Safer == Slower.
// This trick (to change the texID depending on the TLUT addr) is a trick to get around
// an issue with metroid prime's fonts, where it has multiple sets of fonts on top of
// each other stored in a single texture, and uses the palette to make different characters
// visible or invisible. Thus, unless we want to recreate the textures for every drawn character,
// we must make sure that texture with different tluts get different IDs.
u32 tlutHash = TexDecoder_GetTlutHash(&texMem[tlutaddr], (tex_format == GX_TF_C4) ? 32 : 128);
texHash ^= tlutHash;
if (g_ActiveConfig.bSafeTextureCache)
texID ^= tlutHash;
}
}
bool skip_texture_create = false;
TexCache::iterator iter = textures.find(texID);
if (iter != textures.end())
{
TCacheEntry &entry = iter->second;
if (!g_ActiveConfig.bSafeTextureCache)
hash_value = ((u32 *)ptr)[0];
if (entry.isRenderTarget || ((address == entry.addr) && (hash_value == entry.hash)))
{
entry.frameCount = frameCount;
D3D::SetTexture(stage, entry.texture);
return &entry;
}
else
{
// Let's reload the new texture data into the same texture,
// instead of destroying it and having to create a new one.
// Might speed up movie playback very, very slightly.
if (width == entry.w && height==entry.h &&(tex_format | (tlutfmt << 16)) == entry.fmt)
{
skip_texture_create = true;
}
else
{
entry.Destroy(false);
textures.erase(iter);
}
}
}
PC_TexFormat pcfmt = TexDecoder_Decode(temp, ptr, expandedWidth, height, tex_format, tlutaddr, tlutfmt);
D3DFORMAT d3d_fmt;
@ -255,7 +411,7 @@ TextureCache::TCacheEntry *TextureCache::Load(int stage, u32 address, int width,
entry.frameCount = frameCount;
entry.w = width;
entry.h = height;
entry.fmt = tex_format;
entry.fmt = tex_format | (tlutfmt << 16);
if (g_ActiveConfig.bDumpTextures)
{
@ -287,7 +443,6 @@ TextureCache::TCacheEntry *TextureCache::Load(int stage, u32 address, int width,
DEBUGGER_PAUSE_LOG_AT(NEXT_NEW_TEXTURE,true,{printf("A new texture (%d x %d) is loaded", width, height);});
return &entry;
}
void TextureCache::CopyRenderTargetToTexture(u32 address, bool bFromZBuffer, bool bIsIntensityFmt, u32 copyfmt, int bScaleByHalf, const EFBRectangle &source_rect)
{
int efb_w = source_rect.GetWidth();

View File

@ -41,6 +41,8 @@ public:
int frameCount;
int w, h, fmt;
float scaleX, scaleY; // Hires texutres need this
bool isRenderTarget;
bool isNonPow2;
@ -51,6 +53,12 @@ public:
hash = 0;
paletteHash = 0;
oldpixel = 0;
addr = 0;
size_in_bytes = 0;
frameCount = 0;
scaleX = 1.0f;
scaleY = 1.0f;
isNonPow2 = true;
}
void Destroy(bool shutdown);
bool IntersectsMemoryRange(u32 range_address, u32 range_size);

View File

@ -363,8 +363,8 @@ void EncodeToRam(u32 address, bool bFromZBuffer, bool bIsIntensityFmt, u32 copyf
TextureConversionShader::SetShaderParameters(
(float)expandedWidth,
expandedHeight * MValueY,
ceilf(source.left * MValueX + 1.0f),
ceilf(source.top * MValueY + 1.0f),
source.left * MValueX,
source.top * MValueY,
sampleStride * MValueX,
sampleStride * MValueY,
(float)Renderer::GetTargetWidth(),

View File

@ -252,12 +252,15 @@ void Flush()
if (tentry) {
PixelShaderManager::SetTexDims(i, tentry->w, tentry->h, 0, 0);
if (tentry->scaleX != 1.0f || tentry->scaleY != 1.0f)
PixelShaderManager::SetCustomTexScale(i, tentry->scaleX, tentry->scaleY);
}
else
{
DEBUGGER_PAUSE_LOG_AT(NEXT_ERROR,true,{printf("Fail to load texture\n");});
ERROR_LOG(VIDEO, "error loading texture");
}
}
}
PixelShaderManager::SetTexturesUsed(0);

View File

@ -293,13 +293,13 @@ void EncodeToRam(u32 address, bool bFromZBuffer, bool bIsIntensityFmt, u32 copyf
float MValueX = Renderer::GetTargetScaleX();
float MValueY = Renderer::GetTargetScaleY();
float top = Renderer::GetTargetHeight() - floorf((source.top + expandedHeight) * MValueY - 0.5f);
float top = Renderer::GetTargetHeight() - (source.top + expandedHeight) * MValueY ;
float sampleStride = bScaleByHalf?2.0f:1.0f;
TextureConversionShader::SetShaderParameters((float)expandedWidth,
expandedHeight * MValueY,
ceilf(source.left * MValueX + 0.5f),
source.left * MValueX,
top,
sampleStride * MValueX,
sampleStride * MValueY);

View File

@ -310,7 +310,7 @@ TextureMngr::TCacheEntry* TextureMngr::Load(int texstage, u32 address, int width
// Let's reload the new texture data into the same texture,
// instead of destroying it and having to create a new one.
// Might speed up movie playback very, very slightly.
if (width == entry.w && height == entry.h && tex_format == entry.fmt)
if (width == entry.w && height == entry.h && (tex_format | (tlutfmt << 16)) == entry.fmt)
{
glBindTexture(entry.isRectangle ? GL_TEXTURE_RECTANGLE_ARB : GL_TEXTURE_2D, entry.texture);
if (entry.mode.hex != tm0.hex)
@ -463,7 +463,7 @@ TextureMngr::TCacheEntry* TextureMngr::Load(int texstage, u32 address, int width
entry.frameCount = frameCount;
entry.w = width;
entry.h = height;
entry.fmt = tex_format;
entry.fmt = (tex_format | (tlutfmt << 16));
entry.SetTextureParameters(tm0);
if (g_ActiveConfig.bDumpTextures) // dump texture to file
@ -624,6 +624,8 @@ void TextureMngr::CopyRenderTargetToTexture(u32 address, bool bFromZBuffer, bool
int w = (abs(source_rect.GetWidth()) >> bScaleByHalf);
int h = (abs(source_rect.GetHeight()) >> bScaleByHalf);
GL_REPORT_ERRORD();
if (!bIsInit)