diff options
Diffstat (limited to 'src/GPU2D.cpp')
-rw-r--r-- | src/GPU2D.cpp | 2268 |
1 files changed, 40 insertions, 2228 deletions
diff --git a/src/GPU2D.cpp b/src/GPU2D.cpp index 7774c65..fa05e79 100644 --- a/src/GPU2D.cpp +++ b/src/GPU2D.cpp @@ -84,20 +84,6 @@ GPU2D::GPU2D(u32 num) { Num = num; - - // initialize mosaic table - for (int m = 0; m < 16; m++) - { - for (int x = 0; x < 256; x++) - { - int offset = x % (m+1); - MosaicTable[m][x] = offset; - } - } -} - -GPU2D::~GPU2D() -{ } void GPU2D::Reset() @@ -131,8 +117,6 @@ void GPU2D::Reset() BGMosaicYMax = 0; OBJMosaicY = 0; OBJMosaicYMax = 0; - CurBGXMosaicTable = MosaicTable[0]; - CurOBJXMosaicTable = MosaicTable[0]; BlendCnt = 0; EVA = 16; @@ -149,11 +133,7 @@ void GPU2D::Reset() MasterBrightness = 0; - BGExtPalStatus[0] = 0; - BGExtPalStatus[1] = 0; - BGExtPalStatus[2] = 0; - BGExtPalStatus[3] = 0; - OBJExtPalStatus = 0; + MosaicXSizeChanged(); } void GPU2D::DoSavestate(Savestate* file) @@ -206,18 +186,7 @@ void GPU2D::DoSavestate(Savestate* file) file->Var32(&Win0Active); file->Var32(&Win1Active); - if (!file->Saving) - { - // refresh those - BGExtPalStatus[0] = 0; - BGExtPalStatus[1] = 0; - BGExtPalStatus[2] = 0; - BGExtPalStatus[3] = 0; - OBJExtPalStatus = 0; - - CurBGXMosaicTable = MosaicTable[BGMosaicSize[0]]; - CurOBJXMosaicTable = MosaicTable[OBJMosaicSize[0]]; - } + MosaicXSizeChanged(); } void GPU2D::SetFramebuffer(u32* buf) @@ -225,15 +194,6 @@ void GPU2D::SetFramebuffer(u32* buf) Framebuffer = buf; } -void GPU2D::SetRenderSettings(bool accel) -{ - Accelerated = accel; - - if (Accelerated) DrawPixel = DrawPixel_Accel; - else DrawPixel = DrawPixel_Normal; -} - - u8 GPU2D::Read8(u32 addr) { switch (addr & 0x00000FFF) @@ -328,6 +288,13 @@ void GPU2D::Write8(u32 addr, u8 val) DispCnt = (DispCnt & 0x00FFFFFF) | (val << 24); if (Num) DispCnt &= 0xC0B1FFF7; return; + + case 0x10: + if (!Num) GPU3D::SetRenderXPos((GPU3D::RenderXPos & 0xFF00) | val); + break; + case 0x11: + if (!Num) GPU3D::SetRenderXPos((GPU3D::RenderXPos & 0x00FF) | (val << 8)); + break; } if (!Enabled) return; @@ -378,12 +345,12 @@ void GPU2D::Write8(u32 addr, u8 val) case 0x04C: BGMosaicSize[0] = val & 0xF; BGMosaicSize[1] = val >> 4; - CurBGXMosaicTable = MosaicTable[BGMosaicSize[0]]; + MosaicXSizeChanged(); return; case 0x04D: OBJMosaicSize[0] = val & 0xF; OBJMosaicSize[1] = val >> 4; - CurOBJXMosaicTable = MosaicTable[OBJMosaicSize[0]]; + MosaicXSizeChanged(); return; case 0x050: BlendCnt = (BlendCnt & 0x3F00) | val; return; @@ -420,6 +387,10 @@ void GPU2D::Write16(u32 addr, u16 val) if (Num) DispCnt &= 0xC0B1FFF7; return; + case 0x010: + if (!Num) GPU3D::SetRenderXPos(val); + break; + case 0x068: DispFIFO[DispFIFOWritePtr] = val; return; @@ -526,10 +497,9 @@ void GPU2D::Write16(u32 addr, u16 val) case 0x04C: BGMosaicSize[0] = val & 0xF; BGMosaicSize[1] = (val >> 4) & 0xF; - CurBGXMosaicTable = MosaicTable[BGMosaicSize[0]]; OBJMosaicSize[0] = (val >> 8) & 0xF; OBJMosaicSize[1] = val >> 12; - CurOBJXMosaicTable = MosaicTable[OBJMosaicSize[0]]; + MosaicXSizeChanged(); return; case 0x050: BlendCnt = val & 0x3FFF; return; @@ -603,138 +573,6 @@ void GPU2D::Write32(u32 addr, u32 val) Write16(addr+2, val>>16); } - -u32 GPU2D::ColorBlend4(u32 val1, u32 val2, u32 eva, u32 evb) -{ - u32 r = (((val1 & 0x00003F) * eva) + ((val2 & 0x00003F) * evb)) >> 4; - u32 g = ((((val1 & 0x003F00) * eva) + ((val2 & 0x003F00) * evb)) >> 4) & 0x007F00; - u32 b = ((((val1 & 0x3F0000) * eva) + ((val2 & 0x3F0000) * evb)) >> 4) & 0x7F0000; - - if (r > 0x00003F) r = 0x00003F; - if (g > 0x003F00) g = 0x003F00; - if (b > 0x3F0000) b = 0x3F0000; - - return r | g | b | 0xFF000000; -} - -u32 GPU2D::ColorBlend5(u32 val1, u32 val2) -{ - u32 eva = ((val1 >> 24) & 0x1F) + 1; - u32 evb = 32 - eva; - - if (eva == 32) return val1; - - u32 r = (((val1 & 0x00003F) * eva) + ((val2 & 0x00003F) * evb)) >> 5; - u32 g = ((((val1 & 0x003F00) * eva) + ((val2 & 0x003F00) * evb)) >> 5) & 0x007F00; - u32 b = ((((val1 & 0x3F0000) * eva) + ((val2 & 0x3F0000) * evb)) >> 5) & 0x7F0000; - - if (eva <= 16) - { - r += 0x000001; - g += 0x000100; - b += 0x010000; - } - - if (r > 0x00003F) r = 0x00003F; - if (g > 0x003F00) g = 0x003F00; - if (b > 0x3F0000) b = 0x3F0000; - - return r | g | b | 0xFF000000; -} - -u32 GPU2D::ColorBrightnessUp(u32 val, u32 factor) -{ - u32 rb = val & 0x3F003F; - u32 g = val & 0x003F00; - - rb += ((((0x3F003F - rb) * factor) >> 4) & 0x3F003F); - g += ((((0x003F00 - g) * factor) >> 4) & 0x003F00); - - return rb | g | 0xFF000000; -} - -u32 GPU2D::ColorBrightnessDown(u32 val, u32 factor) -{ - u32 rb = val & 0x3F003F; - u32 g = val & 0x003F00; - - rb -= (((rb * factor) >> 4) & 0x3F003F); - g -= (((g * factor) >> 4) & 0x003F00); - - return rb | g | 0xFF000000; -} - -u32 GPU2D::ColorComposite(int i, u32 val1, u32 val2) -{ - u32 coloreffect = 0; - u32 eva, evb; - - u32 flag1 = val1 >> 24; - u32 flag2 = val2 >> 24; - - u32 target2; - if (flag2 & 0x80) target2 = 0x1000; - else if (flag2 & 0x40) target2 = 0x0100; - else target2 = flag2 << 8; - - if ((flag1 & 0x80) && (BlendCnt & target2)) - { - // sprite blending - - coloreffect = 1; - - if (flag1 & 0x40) - { - eva = flag1 & 0x1F; - evb = 16 - eva; - } - else - { - eva = EVA; - evb = EVB; - } - } - else if ((flag1 & 0x40) && (BlendCnt & target2)) - { - // 3D layer blending - - coloreffect = 4; - } - else - { - if (flag1 & 0x80) flag1 = 0x10; - else if (flag1 & 0x40) flag1 = 0x01; - - if ((BlendCnt & flag1) && (WindowMask[i] & 0x20)) - { - coloreffect = (BlendCnt >> 6) & 0x3; - - if (coloreffect == 1) - { - if (BlendCnt & target2) - { - eva = EVA; - evb = EVB; - } - else - coloreffect = 0; - } - } - } - - switch (coloreffect) - { - case 0: return val1; - case 1: return ColorBlend4(val1, val2, eva, evb); - case 2: return ColorBrightnessUp(val1, EVY); - case 3: return ColorBrightnessDown(val1, EVY); - case 4: return ColorBlend5(val1, val2); - } - - return val1; -} - - void GPU2D::UpdateMosaicCounters(u32 line) { // Y mosaic uses incrementing 4-bit counters @@ -752,184 +590,14 @@ void GPU2D::UpdateMosaicCounters(u32 line) } } - -void GPU2D::DrawScanline(u32 line) +void GPU2D::VBlank() { - int stride = Accelerated ? (256*3 + 1) : 256; - u32* dst = &Framebuffer[stride * line]; - - int n3dline = line; - line = GPU::VCount; - - bool forceblank = false; - - // scanlines that end up outside of the GPU drawing range - // (as a result of writing to VCount) are filled white - if (line > 192) forceblank = true; - - // GPU B can be completely disabled by POWCNT1 - // oddly that's not the case for GPU A - if (Num && !Enabled) forceblank = true; - - if (forceblank) - { - for (int i = 0; i < 256; i++) - dst[i] = 0xFFFFFFFF; - - if (Accelerated) - { - dst[256*3] = 0; - } - return; - } - - u32 dispmode = DispCnt >> 16; - dispmode &= (Num ? 0x1 : 0x3); - - if (Num == 0) + if (CaptureLatch) { - if (!Accelerated) - _3DLine = GPU3D::GetLine(n3dline); - else if ((CaptureCnt & (1<<31)) && (((CaptureCnt >> 29) & 0x3) != 1)) - { - _3DLine = GPU3D::GetLine(n3dline); - //GPU3D::GLRenderer::PrepareCaptureFrame(); - } + CaptureCnt &= ~(1<<31); + CaptureLatch = false; } - // always render regular graphics - DrawScanline_BGOBJ(line); - UpdateMosaicCounters(line); - - switch (dispmode) - { - case 0: // screen off - { - for (int i = 0; i < 256; i++) - dst[i] = 0x003F3F3F; - } - break; - - case 1: // regular display - { - int i = 0; - for (; i < (stride & ~1); i+=2) - *(u64*)&dst[i] = *(u64*)&BGOBJLine[i]; - } - break; - - case 2: // VRAM display - { - u32 vrambank = (DispCnt >> 18) & 0x3; - if (GPU::VRAMMap_LCDC & (1<<vrambank)) - { - u16* vram = (u16*)GPU::VRAM[vrambank]; - vram = &vram[line * 256]; - - for (int i = 0; i < 256; i++) - { - u16 color = vram[i]; - u8 r = (color & 0x001F) << 1; - u8 g = (color & 0x03E0) >> 4; - u8 b = (color & 0x7C00) >> 9; - - dst[i] = r | (g << 8) | (b << 16); - } - } - else - { - for (int i = 0; i < 256; i++) - { - dst[i] = 0; - } - } - } - break; - - case 3: // FIFO display - { - for (int i = 0; i < 256; i++) - { - u16 color = DispFIFOBuffer[i]; - u8 r = (color & 0x001F) << 1; - u8 g = (color & 0x03E0) >> 4; - u8 b = (color & 0x7C00) >> 9; - - dst[i] = r | (g << 8) | (b << 16); - } - } - break; - } - - // capture - if ((Num == 0) && (CaptureCnt & (1<<31))) - { - u32 capwidth, capheight; - switch ((CaptureCnt >> 20) & 0x3) - { - case 0: capwidth = 128; capheight = 128; break; - case 1: capwidth = 256; capheight = 64; break; - case 2: capwidth = 256; capheight = 128; break; - case 3: capwidth = 256; capheight = 192; break; - } - - if (line < capheight) - DoCapture(line, capwidth); - } - - if (Accelerated) - { - dst[256*3] = MasterBrightness | (DispCnt & 0x30000); - return; - } - - // master brightness - if (dispmode != 0) - { - if ((MasterBrightness >> 14) == 1) - { - // up - u32 factor = MasterBrightness & 0x1F; - if (factor > 16) factor = 16; - - for (int i = 0; i < 256; i++) - { - dst[i] = ColorBrightnessUp(dst[i], factor); - } - } - else if ((MasterBrightness >> 14) == 2) - { - // down - u32 factor = MasterBrightness & 0x1F; - if (factor > 16) factor = 16; - - for (int i = 0; i < 256; i++) - { - dst[i] = ColorBrightnessDown(dst[i], factor); - } - } - } - - // convert to 32-bit BGRA - // note: 32-bit RGBA would be more straightforward, but - // BGRA seems to be more compatible (Direct2D soft, cairo...) - for (int i = 0; i < 256; i+=2) - { - u64 c = *(u64*)&dst[i]; - - u64 r = (c << 18) & 0xFC000000FC0000; - u64 g = (c << 2) & 0xFC000000FC00; - u64 b = (c >> 14) & 0xFC000000FC; - c = r | g | b; - - *(u64*)&dst[i] = c | ((c & 0x00C0C0C000C0C0C0) >> 6) | 0xFF000000FF000000; - } -} - -void GPU2D::VBlank() -{ - CaptureCnt &= ~(1<<31); - DispFIFOReadPtr = 0; DispFIFOWritePtr = 0; } @@ -948,235 +616,6 @@ void GPU2D::VBlankEnd() //OBJMosaicYMax = OBJMosaicSize[1]; //OBJMosaicY = 0; //OBJMosaicYCount = 0; - -#ifdef OGLRENDERER_ENABLED - if (Accelerated) - { - if ((Num == 0) && (CaptureCnt & (1<<31)) && (((CaptureCnt >> 29) & 0x3) != 1)) - { - GPU3D::GLRenderer::PrepareCaptureFrame(); - } - } -#endif -} - - -void GPU2D::DoCapture(u32 line, u32 width) -{ - u32 dstvram = (CaptureCnt >> 16) & 0x3; - - // TODO: confirm this - // it should work like VRAM display mode, which requires VRAM to be mapped to LCDC - if (!(GPU::VRAMMap_LCDC & (1<<dstvram))) - return; - - u16* dst = (u16*)GPU::VRAM[dstvram]; - u32 dstaddr = (((CaptureCnt >> 18) & 0x3) << 14) + (line * width); - - // TODO: handle 3D in accelerated mode!! - - u32* srcA; - if (CaptureCnt & (1<<24)) - { - srcA = _3DLine; - } - else - { - srcA = BGOBJLine; - if (Accelerated) - { - // in accelerated mode, compositing is normally done on the GPU - // but when doing display capture, we do need the composited output - // so we do it here - - for (int i = 0; i < 256; i++) - { - u32 val1 = BGOBJLine[i]; - u32 val2 = BGOBJLine[256+i]; - u32 val3 = BGOBJLine[512+i]; - - u32 compmode = (val3 >> 24) & 0xF; - - if (compmode == 4) - { - // 3D on top, blending - - u32 _3dval = _3DLine[val3 & 0xFF]; - if ((_3dval >> 24) > 0) - val1 = ColorBlend5(_3dval, val1); - else - val1 = val2; - } - else if (compmode == 1) - { - // 3D on bottom, blending - - u32 _3dval = _3DLine[val3 & 0xFF]; - if ((_3dval >> 24) > 0) - { - u32 eva = (val3 >> 8) & 0x1F; - u32 evb = (val3 >> 16) & 0x1F; - - val1 = ColorBlend4(val1, _3dval, eva, evb); - } - else - val1 = val2; - } - else if (compmode <= 3) - { - // 3D on top, normal/fade - - u32 _3dval = _3DLine[val3 & 0xFF]; - if ((_3dval >> 24) > 0) - { - u32 evy = (val3 >> 8) & 0x1F; - - val1 = _3dval; - if (compmode == 2) val1 = ColorBrightnessUp(val1, evy); - else if (compmode == 3) val1 = ColorBrightnessDown(val1, evy); - } - else - val1 = val2; - } - - BGOBJLine[i] = val1; - } - } - } - - u16* srcB = NULL; - u32 srcBaddr = line * 256; - - if (CaptureCnt & (1<<25)) - { - srcB = &DispFIFOBuffer[0]; - srcBaddr = 0; - } - else - { - u32 srcvram = (DispCnt >> 18) & 0x3; - if (GPU::VRAMMap_LCDC & (1<<srcvram)) - srcB = (u16*)GPU::VRAM[srcvram]; - - if (((DispCnt >> 16) & 0x3) != 2) - srcBaddr += ((CaptureCnt >> 26) & 0x3) << 14; - } - - dstaddr &= 0xFFFF; - srcBaddr &= 0xFFFF; - - switch ((CaptureCnt >> 29) & 0x3) - { - case 0: // source A - { - for (u32 i = 0; i < width; i++) - { - u32 val = srcA[i]; - - // TODO: check what happens when alpha=0 - - u32 r = (val >> 1) & 0x1F; - u32 g = (val >> 9) & 0x1F; - u32 b = (val >> 17) & 0x1F; - u32 a = ((val >> 24) != 0) ? 0x8000 : 0; - - dst[dstaddr] = r | (g << 5) | (b << 10) | a; - dstaddr = (dstaddr + 1) & 0xFFFF; - } - } - break; - - case 1: // source B - { - if (srcB) - { - for (u32 i = 0; i < width; i++) - { - dst[dstaddr] = srcB[srcBaddr]; - srcBaddr = (srcBaddr + 1) & 0xFFFF; - dstaddr = (dstaddr + 1) & 0xFFFF; - } - } - else - { - for (u32 i = 0; i < width; i++) - { - dst[dstaddr] = 0; - dstaddr = (dstaddr + 1) & 0xFFFF; - } - } - } - break; - - case 2: // sources A+B - case 3: - { - u32 eva = CaptureCnt & 0x1F; - u32 evb = (CaptureCnt >> 8) & 0x1F; - - // checkme - if (eva > 16) eva = 16; - if (evb > 16) evb = 16; - - if (srcB) - { - for (u32 i = 0; i < width; i++) - { - u32 val = srcA[i]; - - // TODO: check what happens when alpha=0 - - u32 rA = (val >> 1) & 0x1F; - u32 gA = (val >> 9) & 0x1F; - u32 bA = (val >> 17) & 0x1F; - u32 aA = ((val >> 24) != 0) ? 1 : 0; - - val = srcB[srcBaddr]; - - u32 rB = val & 0x1F; - u32 gB = (val >> 5) & 0x1F; - u32 bB = (val >> 10) & 0x1F; - u32 aB = val >> 15; - - u32 rD = ((rA * aA * eva) + (rB * aB * evb)) >> 4; - u32 gD = ((gA * aA * eva) + (gB * aB * evb)) >> 4; - u32 bD = ((bA * aA * eva) + (bB * aB * evb)) >> 4; - u32 aD = (eva>0 ? aA : 0) | (evb>0 ? aB : 0); - - if (rD > 0x1F) rD = 0x1F; - if (gD > 0x1F) gD = 0x1F; - if (bD > 0x1F) bD = 0x1F; - - dst[dstaddr] = rD | (gD << 5) | (bD << 10) | (aD << 15); - srcBaddr = (srcBaddr + 1) & 0xFFFF; - dstaddr = (dstaddr + 1) & 0xFFFF; - } - } - else - { - for (u32 i = 0; i < width; i++) - { - u32 val = srcA[i]; - - // TODO: check what happens when alpha=0 - - u32 rA = (val >> 1) & 0x1F; - u32 gA = (val >> 9) & 0x1F; - u32 bA = (val >> 17) & 0x1F; - u32 aA = ((val >> 24) != 0) ? 1 : 0; - - u32 rD = (rA * aA * eva) >> 4; - u32 gD = (gA * aA * eva) >> 4; - u32 bD = (bA * aA * eva) >> 4; - u32 aD = (eva>0 ? aA : 0); - - dst[dstaddr] = rD | (gD << 5) | (bD << 10) | (aD << 15); - dstaddr = (dstaddr + 1) & 0xFFFF; - } - } - } - break; - } } void GPU2D::SampleFIFO(u32 offset, u32 num) @@ -1191,88 +630,22 @@ void GPU2D::SampleFIFO(u32 offset, u32 num) } } - -void GPU2D::BGExtPalDirty(u32 base) -{ - BGExtPalStatus[base] = 0; - BGExtPalStatus[base+1] = 0; -} - -void GPU2D::OBJExtPalDirty() -{ - OBJExtPalStatus = 0; -} - - u16* GPU2D::GetBGExtPal(u32 slot, u32 pal) { - u16* dst = &BGExtPalCache[slot][pal << 8]; - - if (!(BGExtPalStatus[slot] & (1<<pal))) - { - if (Num) - { - if (GPU::VRAMMap_BBGExtPal[slot] & (1<<7)) - memcpy(dst, &GPU::VRAM_H[(slot << 13) + (pal << 9)], 256*2); - else - memset(dst, 0, 256*2); - } - else - { - memset(dst, 0, 256*2); - - if (GPU::VRAMMap_ABGExtPal[slot] & (1<<4)) - for (int i = 0; i < 256; i+=2) - *(u32*)&dst[i] |= *(u32*)&GPU::VRAM_E[(slot << 13) + (pal << 9) + (i << 1)]; - - if (GPU::VRAMMap_ABGExtPal[slot] & (1<<5)) - for (int i = 0; i < 256; i+=2) - *(u32*)&dst[i] |= *(u32*)&GPU::VRAM_F[((slot&1) << 13) + (pal << 9) + (i << 1)]; - - if (GPU::VRAMMap_ABGExtPal[slot] & (1<<6)) - for (int i = 0; i < 256; i+=2) - *(u32*)&dst[i] |= *(u32*)&GPU::VRAM_G[((slot&1) << 13) + (pal << 9) + (i << 1)]; - } - - BGExtPalStatus[slot] |= (1<<pal); - } - - return dst; + const u32 PaletteSize = 256 * 2; + const u32 SlotSize = PaletteSize * 16; + return (u16*)&(Num == 0 + ? GPU::VRAMFlat_ABGExtPal + : GPU::VRAMFlat_BBGExtPal)[slot * SlotSize + pal * PaletteSize]; } u16* GPU2D::GetOBJExtPal() { - u16* dst = OBJExtPalCache; - - if (!OBJExtPalStatus) - { - if (Num) - { - if (GPU::VRAMMap_BOBJExtPal & (1<<8)) - memcpy(dst, &GPU::VRAM_I[0], 16*256*2); - else - memset(dst, 0, 16*256*2); - } - else - { - memset(dst, 0, 16*256*2); - - if (GPU::VRAMMap_AOBJExtPal & (1<<5)) - for (int i = 0; i < 16*256; i+=2) - *(u32*)&dst[i] |= *(u32*)&GPU::VRAM_F[i << 1]; - - if (GPU::VRAMMap_AOBJExtPal & (1<<6)) - for (int i = 0; i < 16*256; i+=2) - *(u32*)&dst[i] |= *(u32*)&GPU::VRAM_G[i << 1]; - } - - OBJExtPalStatus = 1; - } - - return dst; + return Num == 0 + ? (u16*)GPU::VRAMFlat_AOBJExtPal + : (u16*)GPU::VRAMFlat_BOBJExtPal; } - void GPU2D::CheckWindows(u32 line) { line &= 0xFF; @@ -1328,1591 +701,30 @@ void GPU2D::CalculateWindowMask(u32 line) } } - -#define DoDrawBG(type, line, num) \ - { if ((BGCnt[num] & 0x0040) && (BGMosaicSize[0] > 0)) DrawBG_##type<true>(line, num); else DrawBG_##type<false>(line, num); } - -#define DoDrawBG_Large(line) \ - { if ((BGCnt[2] & 0x0040) && (BGMosaicSize[0] > 0)) DrawBG_Large<true>(line); else DrawBG_Large<false>(line); } - -template<u32 bgmode> -void GPU2D::DrawScanlineBGMode(u32 line) -{ - for (int i = 3; i >= 0; i--) - { - if ((BGCnt[3] & 0x3) == i) - { - if (DispCnt & 0x0800) - { - if (bgmode >= 3) - DoDrawBG(Extended, line, 3) - else if (bgmode >= 1) - DoDrawBG(Affine, line, 3) - else - DoDrawBG(Text, line, 3) - } - } - if ((BGCnt[2] & 0x3) == i) - { - if (DispCnt & 0x0400) - { - if (bgmode == 5) - DoDrawBG(Extended, line, 2) - else if (bgmode == 4 || bgmode == 2) - DoDrawBG(Affine, line, 2) - else - DoDrawBG(Text, line, 2) - } - } - if ((BGCnt[1] & 0x3) == i) - { - if (DispCnt & 0x0200) - { - DoDrawBG(Text, line, 1) - } - } - if ((BGCnt[0] & 0x3) == i) - { - if (DispCnt & 0x0100) - { - if ((!Num) && (DispCnt & 0x8)) - DrawBG_3D(); - else - DoDrawBG(Text, line, 0) - } - } - if ((DispCnt & 0x1000) && NumSprites) - InterleaveSprites(0x40000 | (i<<16)); - } -} - -void GPU2D::DrawScanlineBGMode6(u32 line) -{ - for (int i = 3; i >= 0; i--) - { - if ((BGCnt[2] & 0x3) == i) - { - if (DispCnt & 0x0400) - { - DoDrawBG_Large(line) - } - } - if ((BGCnt[0] & 0x3) == i) - { - if (DispCnt & 0x0100) - { - if ((!Num) && (DispCnt & 0x8)) - DrawBG_3D(); - } - } - if ((DispCnt & 0x1000) && NumSprites) - InterleaveSprites(0x40000 | (i<<16)); - } -} - -void GPU2D::DrawScanlineBGMode7(u32 line) -{ - // mode 7 only has text-mode BG0 and BG1 - - for (int i = 3; i >= 0; i--) - { - if ((BGCnt[1] & 0x3) == i) - { - if (DispCnt & 0x0200) - { - DoDrawBG(Text, line, 1) - } - } - if ((BGCnt[0] & 0x3) == i) - { - if (DispCnt & 0x0100) - { - if ((!Num) && (DispCnt & 0x8)) - DrawBG_3D(); - else - DoDrawBG(Text, line, 0) - } - } - if ((DispCnt & 0x1000) && NumSprites) - InterleaveSprites(0x40000 | (i<<16)); - } -} - -void GPU2D::DrawScanline_BGOBJ(u32 line) +void GPU2D::GetBGVRAM(u8*& data, u32& mask) { - // forced blank disables BG/OBJ compositing - if (DispCnt & (1<<7)) - { - for (int i = 0; i < 256; i++) - BGOBJLine[i] = 0xFF3F3F3F; - - return; - } - - u64 backdrop; - if (Num) backdrop = *(u16*)&GPU::Palette[0x400]; - else backdrop = *(u16*)&GPU::Palette[0]; - - { - u8 r = (backdrop & 0x001F) << 1; - u8 g = (backdrop & 0x03E0) >> 4; - u8 b = (backdrop & 0x7C00) >> 9; - - backdrop = r | (g << 8) | (b << 16) | 0x20000000; - backdrop |= (backdrop << 32); - - for (int i = 0; i < 256; i+=2) - *(u64*)&BGOBJLine[i] = backdrop; - } - - if (DispCnt & 0xE000) - CalculateWindowMask(line); - else - memset(WindowMask, 0xFF, 256); - - ApplySpriteMosaicX(); - - switch (DispCnt & 0x7) - { - case 0: DrawScanlineBGMode<0>(line); break; - case 1: DrawScanlineBGMode<1>(line); break; - case 2: DrawScanlineBGMode<2>(line); break; - case 3: DrawScanlineBGMode<3>(line); break; - case 4: DrawScanlineBGMode<4>(line); break; - case 5: DrawScanlineBGMode<5>(line); break; - case 6: DrawScanlineBGMode6(line); break; - case 7: DrawScanlineBGMode7(line); break; - } - - // color special effects - // can likely be optimized - - if (!Accelerated) - { - for (int i = 0; i < 256; i++) - { - u32 val1 = BGOBJLine[i]; - u32 val2 = BGOBJLine[256+i]; - - BGOBJLine[i] = ColorComposite(i, val1, val2); - } - } - else - { - if (Num == 0) - { - for (int i = 0; i < 256; i++) - { - u32 val1 = BGOBJLine[i]; - u32 val2 = BGOBJLine[256+i]; - u32 val3 = BGOBJLine[512+i]; - - u32 flag1 = val1 >> 24; - u32 flag2 = val2 >> 24; - - u32 bldcnteffect = (BlendCnt >> 6) & 0x3; - - u32 target1; - if (flag1 & 0x80) target1 = 0x0010; - else if (flag1 & 0x40) target1 = 0x0001; - else target1 = flag1; - - u32 target2; - if (flag2 & 0x80) target2 = 0x1000; - else if (flag2 & 0x40) target2 = 0x0100; - else target2 = flag2 << 8; - - if (((flag1 & 0xC0) == 0x40) && (BlendCnt & target2)) - { - // 3D on top, blending - - BGOBJLine[i] = val2; - BGOBJLine[256+i] = ColorComposite(i, val2, val3); - BGOBJLine[512+i] = 0x04000000 | (val1 & 0xFF); - } - else if ((flag1 & 0xC0) == 0x40) - { - // 3D on top, normal/fade - - if (bldcnteffect == 1) bldcnteffect = 0; - if (!(BlendCnt & 0x0001)) bldcnteffect = 0; - if (!(WindowMask[i] & 0x20)) bldcnteffect = 0; - - BGOBJLine[i] = val2; - BGOBJLine[256+i] = ColorComposite(i, val2, val3); - BGOBJLine[512+i] = (bldcnteffect << 24) | (EVY << 8) | (val1 & 0xFF); - } - else if (((flag2 & 0xC0) == 0x40) && ((BlendCnt & 0x01C0) == 0x0140)) - { - // 3D on bottom, blending - - u32 eva, evb; - if ((flag1 & 0xC0) == 0xC0) - { - eva = flag1 & 0x1F; - evb = 16 - eva; - } - else if (((BlendCnt & target1) && (WindowMask[i] & 0x20)) || - ((flag1 & 0xC0) == 0x80)) - { - eva = EVA; - evb = EVB; - } - else - bldcnteffect = 7; - - BGOBJLine[i] = val1; - BGOBJLine[256+i] = ColorComposite(i, val1, val3); - BGOBJLine[512+i] = (bldcnteffect << 24) | (EVB << 16) | (EVA << 8) | (val2 & 0xFF); - } - else - { - // no potential 3D pixel involved - - BGOBJLine[i] = ColorComposite(i, val1, val2); - BGOBJLine[256+i] = 0; - BGOBJLine[512+i] = 0x07000000; - } - } - } - else - { - for (int i = 0; i < 256; i++) - { - u32 val1 = BGOBJLine[i]; - u32 val2 = BGOBJLine[256+i]; - - BGOBJLine[i] = ColorComposite(i, val1, val2); - BGOBJLine[256+i] = 0; - BGOBJLine[512+i] = 0x07000000; - } - } - } - - if (BGMosaicY >= BGMosaicYMax) - { - BGMosaicY = 0; - BGMosaicYMax = BGMosaicSize[1]; - } - else - BGMosaicY++; - - /*if (OBJMosaicY >= OBJMosaicYMax) - { - OBJMosaicY = 0; - OBJMosaicYMax = OBJMosaicSize[1]; - } - else - OBJMosaicY++;*/ -} - - -void GPU2D::DrawPixel_Normal(u32* dst, u16 color, u32 flag) -{ - u8 r = (color & 0x001F) << 1; - u8 g = (color & 0x03E0) >> 4; - u8 b = (color & 0x7C00) >> 9; - //g |= ((color & 0x8000) >> 15); - - *(dst+256) = *dst; - *dst = r | (g << 8) | (b << 16) | flag; -} - -void GPU2D::DrawPixel_Accel(u32* dst, u16 color, u32 flag) -{ - u8 r = (color & 0x001F) << 1; - u8 g = (color & 0x03E0) >> 4; - u8 b = (color & 0x7C00) >> 9; - - *(dst+512) = *(dst+256); - *(dst+256) = *dst; - *dst = r | (g << 8) | (b << 16) | flag; -} - -void GPU2D::DrawBG_3D() -{ - u16 xoff = BGXPos[0]; - int i = 0; - int iend = 256; - - if (xoff & 0x100) - { - i = (0x100 - (xoff & 0xFF)); - xoff += i; - } - if ((xoff - i + iend - 1) & 0x100) - { - iend -= (xoff & 0xFF); - } - - if (Accelerated) - { - for (; i < iend; i++) - { - int pos = xoff++; - - if (!(WindowMask[i] & 0x01)) continue; - - BGOBJLine[i+512] = BGOBJLine[i+256]; - BGOBJLine[i+256] = BGOBJLine[i]; - BGOBJLine[i] = 0x40000000 | pos; // 3D-layer placeholder - } - } - else - { - for (; i < iend; i++) - { - u32 c = _3DLine[xoff]; - xoff++; - - if ((c >> 24) == 0) continue; - if (!(WindowMask[i] & 0x01)) continue; - - BGOBJLine[i+256] = BGOBJLine[i]; - BGOBJLine[i] = c | 0x40000000; - } - } -} - -template<bool mosaic> -void GPU2D::DrawBG_Text(u32 line, u32 bgnum) -{ - u16 bgcnt = BGCnt[bgnum]; - - u32 tilesetaddr, tilemapaddr; - u16* pal; - u32 extpal, extpalslot; - - u16 xoff = BGXPos[bgnum]; - u16 yoff = BGYPos[bgnum] + line; - - if (bgcnt & 0x0040) - { - // vertical mosaic - yoff -= BGMosaicY; - } - - u32 widexmask = (bgcnt & 0x4000) ? 0x100 : 0; - - extpal = (DispCnt & 0x40000000); - if (extpal) extpalslot = ((bgnum<2) && (bgcnt&0x2000)) ? (2+bgnum) : bgnum; - - if (Num) - { - tilesetaddr = 0x06200000 + ((bgcnt & 0x003C) << 12); - tilemapaddr = 0x06200000 + ((bgcnt & 0x1F00) << 3); - - pal = (u16*)&GPU::Palette[0x400]; - } - else - { - tilesetaddr = 0x06000000 + ((DispCnt & 0x07000000) >> 8) + ((bgcnt & 0x003C) << 12); - tilemapaddr = 0x06000000 + ((DispCnt & 0x38000000) >> 11) + ((bgcnt & 0x1F00) << 3); - - pal = (u16*)&GPU::Palette[0]; - } - - // adjust Y position in tilemap - if (bgcnt & 0x8000) - { - tilemapaddr += ((yoff & 0x1F8) << 3); - if (bgcnt & 0x4000) - tilemapaddr += ((yoff & 0x100) << 3); - } - else - tilemapaddr += ((yoff & 0xF8) << 3); - - u16 curtile; - u16* curpal; - u32 pixelsaddr; - u8 color; - u32 lastxpos; - - if (bgcnt & 0x0080) - { - // 256-color - - // preload shit as needed - if ((xoff & 0x7) || mosaic) - { - curtile = GPU::ReadVRAM_BG<u16>(tilemapaddr + ((xoff & 0xF8) >> 2) + ((xoff & widexmask) << 3)); - - if (extpal) curpal = GetBGExtPal(extpalslot, curtile>>12); - else curpal = pal; - - pixelsaddr = tilesetaddr + ((curtile & 0x03FF) << 6) - + (((curtile & 0x0800) ? (7-(yoff&0x7)) : (yoff&0x7)) << 3); - } - - if (mosaic) lastxpos = xoff; - - for (int i = 0; i < 256; i++) - { - u32 xpos; - if (mosaic) xpos = xoff - CurBGXMosaicTable[i]; - else xpos = xoff; - - if ((!mosaic && (!(xpos & 0x7))) || - (mosaic && ((xpos >> 3) != (lastxpos >> 3)))) - { - // load a new tile - curtile = GPU::ReadVRAM_BG<u16>(tilemapaddr + ((xpos & 0xF8) >> 2) + ((xpos & widexmask) << 3)); - - if (extpal) curpal = GetBGExtPal(extpalslot, curtile>>12); - else curpal = pal; - - pixelsaddr = tilesetaddr + ((curtile & 0x03FF) << 6) - + (((curtile & 0x0800) ? (7-(yoff&0x7)) : (yoff&0x7)) << 3); - - if (mosaic) lastxpos = xpos; - } - - // draw pixel - if (WindowMask[i] & (1<<bgnum)) - { - u32 tilexoff = (curtile & 0x0400) ? (7-(xpos&0x7)) : (xpos&0x7); - color = GPU::ReadVRAM_BG<u8>(pixelsaddr + tilexoff); - - if (color) - DrawPixel(&BGOBJLine[i], curpal[color], 0x01000000<<bgnum); - } - - xoff++; - } - } - else - { - // 16-color - - // preload shit as needed - if ((xoff & 0x7) || mosaic) - { - curtile = GPU::ReadVRAM_BG<u16>(tilemapaddr + ((xoff & 0xF8) >> 2) + ((xoff & widexmask) << 3)); - curpal = pal + ((curtile & 0xF000) >> 8); - pixelsaddr = tilesetaddr + ((curtile & 0x03FF) << 5) - + (((curtile & 0x0800) ? (7-(yoff&0x7)) : (yoff&0x7)) << 2); - } - - if (mosaic) lastxpos = xoff; - - for (int i = 0; i < 256; i++) - { - u32 xpos; - if (mosaic) xpos = xoff - CurBGXMosaicTable[i]; - else xpos = xoff; - - if ((!mosaic && (!(xpos & 0x7))) || - (mosaic && ((xpos >> 3) != (lastxpos >> 3)))) - { - // load a new tile - curtile = GPU::ReadVRAM_BG<u16>(tilemapaddr + ((xpos & 0xF8) >> 2) + ((xpos & widexmask) << 3)); - curpal = pal + ((curtile & 0xF000) >> 8); - pixelsaddr = tilesetaddr + ((curtile & 0x03FF) << 5) - + (((curtile & 0x0800) ? (7-(yoff&0x7)) : (yoff&0x7)) << 2); - - if (mosaic) lastxpos = xpos; - } - - // draw pixel - if (WindowMask[i] & (1<<bgnum)) - { - u32 tilexoff = (curtile & 0x0400) ? (7-(xpos&0x7)) : (xpos&0x7); - if (tilexoff & 0x1) - { - color = GPU::ReadVRAM_BG<u8>(pixelsaddr + (tilexoff >> 1)) >> 4; - } - else - { - color = GPU::ReadVRAM_BG<u8>(pixelsaddr + (tilexoff >> 1)) & 0x0F; - } - - if (color) - DrawPixel(&BGOBJLine[i], curpal[color], 0x01000000<<bgnum); - } - - xoff++; - } - } -} - -template<bool mosaic> -void GPU2D::DrawBG_Affine(u32 line, u32 bgnum) -{ - u16 bgcnt = BGCnt[bgnum]; - - u32 tilesetaddr, tilemapaddr; - u16* pal; - - u32 coordmask; - u32 yshift; - switch (bgcnt & 0xC000) - { - case 0x0000: coordmask = 0x07800; yshift = 7; break; - case 0x4000: coordmask = 0x0F800; yshift = 8; break; - case 0x8000: coordmask = 0x1F800; yshift = 9; break; - case 0xC000: coordmask = 0x3F800; yshift = 10; break; - } - - u32 overflowmask; - if (bgcnt & 0x2000) overflowmask = 0; - else overflowmask = ~(coordmask | 0x7FF); - - s16 rotA = BGRotA[bgnum-2]; - s16 rotB = BGRotB[bgnum-2]; - s16 rotC = BGRotC[bgnum-2]; - s16 rotD = BGRotD[bgnum-2]; - - s32 rotX = BGXRefInternal[bgnum-2]; - s32 rotY = BGYRefInternal[bgnum-2]; - - if (bgcnt & 0x0040) - { - // vertical mosaic - rotX -= (BGMosaicY * rotB); - rotY -= (BGMosaicY * rotD); - } - - if (Num) - { - tilesetaddr = 0x06200000 + ((bgcnt & 0x003C) << 12); - tilemapaddr = 0x06200000 + ((bgcnt & 0x1F00) << 3); - - pal = (u16*)&GPU::Palette[0x400]; - } - else - { - tilesetaddr = 0x06000000 + ((DispCnt & 0x07000000) >> 8) + ((bgcnt & 0x003C) << 12); - tilemapaddr = 0x06000000 + ((DispCnt & 0x38000000) >> 11) + ((bgcnt & 0x1F00) << 3); - - pal = (u16*)&GPU::Palette[0]; - } - - u16 curtile; - u8 color; - - yshift -= 3; - - for (int i = 0; i < 256; i++) - { - if (WindowMask[i] & (1<<bgnum)) - { - s32 finalX, finalY; - if (mosaic) - { - int im = CurBGXMosaicTable[i]; - finalX = rotX - (im * rotA); - finalY = rotY - (im * rotC); - } - else - { - finalX = rotX; - finalY = rotY; - } - - if ((!((finalX|finalY) & overflowmask))) - { - curtile = GPU::ReadVRAM_BG<u8>(tilemapaddr + ((((finalY & coordmask) >> 11) << yshift) + ((finalX & coordmask) >> 11))); - - // draw pixel - u32 tilexoff = (finalX >> 8) & 0x7; - u32 tileyoff = (finalY >> 8) & 0x7; - - color = GPU::ReadVRAM_BG<u8>(tilesetaddr + (curtile << 6) + (tileyoff << 3) + tilexoff); - - if (color) - DrawPixel(&BGOBJLine[i], pal[color], 0x01000000<<bgnum); - } - } - - rotX += rotA; - rotY += rotC; - } - - BGXRefInternal[bgnum-2] += rotB; - BGYRefInternal[bgnum-2] += rotD; -} - -template<bool mosaic> -void GPU2D::DrawBG_Extended(u32 line, u32 bgnum) -{ - u16 bgcnt = BGCnt[bgnum]; - - u32 tilesetaddr, tilemapaddr; - u16* pal; - u32 extpal; - - extpal = (DispCnt & 0x40000000); - - s16 rotA = BGRotA[bgnum-2]; - s16 rotB = BGRotB[bgnum-2]; - s16 rotC = BGRotC[bgnum-2]; - s16 rotD = BGRotD[bgnum-2]; - - s32 rotX = BGXRefInternal[bgnum-2]; - s32 rotY = BGYRefInternal[bgnum-2]; - - if (bgcnt & 0x0040) - { - // vertical mosaic - rotX -= (BGMosaicY * rotB); - rotY -= (BGMosaicY * rotD); - } - - if (bgcnt & 0x0080) - { - // bitmap modes - - u32 xmask, ymask; - u32 yshift; - switch (bgcnt & 0xC000) - { - case 0x0000: xmask = 0x07FFF; ymask = 0x07FFF; yshift = 7; break; - case 0x4000: xmask = 0x0FFFF; ymask = 0x0FFFF; yshift = 8; break; - case 0x8000: xmask = 0x1FFFF; ymask = 0x0FFFF; yshift = 9; break; - case 0xC000: xmask = 0x1FFFF; ymask = 0x1FFFF; yshift = 9; break; - } - - u32 ofxmask, ofymask; - if (bgcnt & 0x2000) - { - ofxmask = 0; - ofymask = 0; - } - else - { - ofxmask = ~xmask; - ofymask = ~ymask; - } - - if (Num) tilemapaddr = 0x06200000 + ((bgcnt & 0x1F00) << 6); - else tilemapaddr = 0x06000000 + ((bgcnt & 0x1F00) << 6); - - if (bgcnt & 0x0004) - { - // direct color bitmap - - u16 color; - - for (int i = 0; i < 256; i++) - { - if (WindowMask[i] & (1<<bgnum)) - { - s32 finalX, finalY; - if (mosaic) - { - int im = CurBGXMosaicTable[i]; - finalX = rotX - (im * rotA); - finalY = rotY - (im * rotC); - } - else - { - finalX = rotX; - finalY = rotY; - } - - if (!(finalX & ofxmask) && !(finalY & ofymask)) - { - color = GPU::ReadVRAM_BG<u16>(tilemapaddr + (((((finalY & ymask) >> 8) << yshift) + ((finalX & xmask) >> 8)) << 1)); - - if (color & 0x8000) - DrawPixel(&BGOBJLine[i], color, 0x01000000<<bgnum); - } - } - - rotX += rotA; - rotY += rotC; - } - } - else - { - // 256-color bitmap - - if (Num) pal = (u16*)&GPU::Palette[0x400]; - else pal = (u16*)&GPU::Palette[0]; - - u8 color; - - for (int i = 0; i < 256; i++) - { - if (WindowMask[i] & (1<<bgnum)) - { - s32 finalX, finalY; - if (mosaic) - { - int im = CurBGXMosaicTable[i]; - finalX = rotX - (im * rotA); - finalY = rotY - (im * rotC); - } - else - { - finalX = rotX; - finalY = rotY; - } - - if (!(finalX & ofxmask) && !(finalY & ofymask)) - { - color = GPU::ReadVRAM_BG<u8>(tilemapaddr + (((finalY & ymask) >> 8) << yshift) + ((finalX & xmask) >> 8)); - - if (color) - DrawPixel(&BGOBJLine[i], pal[color], 0x01000000<<bgnum); - } - } - - rotX += rotA; - rotY += rotC; - } - } - } - else - { - // mixed affine/text mode - - u32 coordmask; - u32 yshift; - switch (bgcnt & 0xC000) - { - case 0x0000: coordmask = 0x07800; yshift = 7; break; - case 0x4000: coordmask = 0x0F800; yshift = 8; break; - case 0x8000: coordmask = 0x1F800; yshift = 9; break; - case 0xC000: coordmask = 0x3F800; yshift = 10; break; - } - - u32 overflowmask; - if (bgcnt & 0x2000) overflowmask = 0; - else overflowmask = ~(coordmask | 0x7FF); - - if (Num) - { - tilesetaddr = 0x06200000 + ((bgcnt & 0x003C) << 12); - tilemapaddr = 0x06200000 + ((bgcnt & 0x1F00) << 3); - - pal = (u16*)&GPU::Palette[0x400]; - } - else - { - tilesetaddr = 0x06000000 + ((DispCnt & 0x07000000) >> 8) + ((bgcnt & 0x003C) << 12); - tilemapaddr = 0x06000000 + ((DispCnt & 0x38000000) >> 11) + ((bgcnt & 0x1F00) << 3); - - pal = (u16*)&GPU::Palette[0]; - } - - u16 curtile; - u16* curpal; - u8 color; - - yshift -= 3; - - for (int i = 0; i < 256; i++) - { - if (WindowMask[i] & (1<<bgnum)) - { - s32 finalX, finalY; - if (mosaic) - { - int im = CurBGXMosaicTable[i]; - finalX = rotX - (im * rotA); - finalY = rotY - (im * rotC); - } - else - { - finalX = rotX; - finalY = rotY; - } - - if ((!((finalX|finalY) & overflowmask))) - { - curtile = GPU::ReadVRAM_BG<u16>(tilemapaddr + (((((finalY & coordmask) >> 11) << yshift) + ((finalX & coordmask) >> 11)) << 1)); - - if (extpal) curpal = GetBGExtPal(bgnum, curtile>>12); - else curpal = pal; - - // draw pixel - u32 tilexoff = (finalX >> 8) & 0x7; - u32 tileyoff = (finalY >> 8) & 0x7; - - if (curtile & 0x0400) tilexoff = 7-tilexoff; - if (curtile & 0x0800) tileyoff = 7-tileyoff; - - color = GPU::ReadVRAM_BG<u8>(tilesetaddr + ((curtile & 0x03FF) << 6) + (tileyoff << 3) + tilexoff); - - if (color) - DrawPixel(&BGOBJLine[i], curpal[color], 0x01000000<<bgnum); - } - } - - rotX += rotA; - rotY += rotC; - } - } - - BGXRefInternal[bgnum-2] += rotB; - BGYRefInternal[bgnum-2] += rotD; -} - -template<bool mosaic> -void GPU2D::DrawBG_Large(u32 line) // BG is always BG2 -{ - u16 bgcnt = BGCnt[2]; - - u32 tilesetaddr, tilemapaddr; - u16* pal; - - // large BG sizes: - // 0: 512x1024 - // 1: 1024x512 - // 2: 512x256 - // 3: 512x512 - u32 xmask, ymask; - u32 yshift; - switch (bgcnt & 0xC000) - { - case 0x0000: xmask = 0x1FFFF; ymask = 0x3FFFF; yshift = 9; break; - case 0x4000: xmask = 0x3FFFF; ymask = 0x1FFFF; yshift = 10; break; - case 0x8000: xmask = 0x1FFFF; ymask = 0x0FFFF; yshift = 9; break; - case 0xC000: xmask = 0x1FFFF; ymask = 0x1FFFF; yshift = 9; break; - } - - u32 ofxmask, ofymask; - if (bgcnt & 0x2000) - { - ofxmask = 0; - ofymask = 0; - } - else - { - ofxmask = ~xmask; - ofymask = ~ymask; - } - - s16 rotA = BGRotA[0]; - s16 rotB = BGRotB[0]; - s16 rotC = BGRotC[0]; - s16 rotD = BGRotD[0]; - - s32 rotX = BGXRefInternal[0]; - s32 rotY = BGYRefInternal[0]; - - if (bgcnt & 0x0040) - { - // vertical mosaic - rotX -= (BGMosaicY * rotB); - rotY -= (BGMosaicY * rotD); - } - - if (Num) tilemapaddr = 0x06200000; - else tilemapaddr = 0x06000000; - - // 256-color bitmap - - if (Num) pal = (u16*)&GPU::Palette[0x400]; - else pal = (u16*)&GPU::Palette[0]; - - u8 color; - - for (int i = 0; i < 256; i++) - { - if (WindowMask[i] & (1<<2)) - { - s32 finalX, finalY; - if (mosaic) - { - int im = CurBGXMosaicTable[i]; - finalX = rotX - (im * rotA); - finalY = rotY - (im * rotC); - } - else - { - finalX = rotX; - finalY = rotY; - } - - if (!(finalX & ofxmask) && !(finalY & ofymask)) - { - color = GPU::ReadVRAM_BG<u8>(tilemapaddr + (((finalY & ymask) >> 8) << yshift) + ((finalX & xmask) >> 8)); - - if (color) - DrawPixel(&BGOBJLine[i], pal[color], 0x01000000<<2); - } - } - - rotX += rotA; - rotY += rotC; - } - - BGXRefInternal[0] += rotB; - BGYRefInternal[0] += rotD; -} - -// OBJ line buffer: -// * bit0-15: color (bit15=1: direct color, bit15=0: palette index, bit12=0 to indicate extpal) -// * bit16-17: BG-relative priority -// * bit18: non-transparent sprite pixel exists here -// * bit19: X mosaic should be applied here -// * bit24-31: compositor flags - -void GPU2D::ApplySpriteMosaicX() -{ - // apply X mosaic if needed - // X mosaic for sprites is applied after all sprites are rendered - - if (OBJMosaicSize[0] == 0) return; - - u32 lastcolor = OBJLine[0]; - - for (u32 i = 1; i < 256; i++) - { - if (!(OBJLine[i] & 0x100000)) - { - // not a mosaic'd sprite pixel - continue; - } - - if ((OBJIndex[i] != OBJIndex[i-1]) || (CurOBJXMosaicTable[i] == 0)) - lastcolor = OBJLine[i]; - else - OBJLine[i] = lastcolor; - } -} - -void GPU2D::InterleaveSprites(u32 prio) -{ - u16* pal = (u16*)&GPU::Palette[Num ? 0x600 : 0x200]; - - if (DispCnt & 0x80000000) - { - u16* extpal = GetOBJExtPal(); - - for (u32 i = 0; i < 256; i++) - { - if ((OBJLine[i] & 0x70000) != prio) continue; - if (!(WindowMask[i] & 0x10)) continue; - - u16 color; - u32 pixel = OBJLine[i]; - - if (pixel & 0x8000) - color = pixel & 0x7FFF; - else if (pixel & 0x1000) - color = pal[pixel & 0xFF]; - else - color = extpal[pixel & 0xFFF]; - - DrawPixel(&BGOBJLine[i], color, pixel & 0xFF000000); - } - } - else - { - // optimized no-extpal version - - for (u32 i = 0; i < 256; i++) - { - if ((OBJLine[i] & 0x70000) != prio) continue; - if (!(WindowMask[i] & 0x10)) continue; - - u16 color; - u32 pixel = OBJLine[i]; - - if (pixel & 0x8000) - color = pixel & 0x7FFF; - else - color = pal[pixel & 0xFF]; - - DrawPixel(&BGOBJLine[i], color, pixel & 0xFF000000); - } - } -} - -#define DoDrawSprite(type, ...) \ - if (iswin) \ - { \ - DrawSprite_##type<true>(__VA_ARGS__); \ - } \ - else \ - { \ - DrawSprite_##type<false>(__VA_ARGS__); \ - } - -void GPU2D::DrawSprites(u32 line) -{ - if (line == 0) - { - // reset those counters here - // TODO: find out when those are supposed to be reset - // it would make sense to reset them at the end of VBlank - // however, sprites are rendered one scanline in advance - // so they need to be reset a bit earlier - - OBJMosaicY = 0; - OBJMosaicYCount = 0; - } - - NumSprites = 0; - memset(OBJLine, 0, 256*4); - memset(OBJWindow, 0, 256); - if (!(DispCnt & 0x1000)) return; - - memset(OBJIndex, 0xFF, 256); - - u16* oam = (u16*)&GPU::OAM[Num ? 0x400 : 0]; - - const s32 spritewidth[16] = - { - 8, 16, 8, 8, - 16, 32, 8, 8, - 32, 32, 16, 8, - 64, 64, 32, 8 - }; - const s32 spriteheight[16] = - { - 8, 8, 16, 8, - 16, 8, 32, 8, - 32, 16, 32, 8, - 64, 32, 64, 8 - }; - - for (int bgnum = 0x0C00; bgnum >= 0x0000; bgnum -= 0x0400) - { - for (int sprnum = 127; sprnum >= 0; sprnum--) - { - u16* attrib = &oam[sprnum*4]; - - if ((attrib[2] & 0x0C00) != bgnum) - continue; - - bool iswin = (((attrib[0] >> 10) & 0x3) == 2); - - u32 sprline; - if ((attrib[0] & 0x1000) && !iswin) - { - // apply Y mosaic - sprline = OBJMosaicY; - } - else - sprline = line; - - if (attrib[0] & 0x0100) - { - u32 sizeparam = (attrib[0] >> 14) | ((attrib[1] & 0xC000) >> 12); - s32 width = spritewidth[sizeparam]; - s32 height = spriteheight[sizeparam]; - s32 boundwidth = width; - s32 boundheight = height; - - if (attrib[0] & 0x0200) - { - boundwidth <<= 1; - boundheight <<= 1; - } - - u32 ypos = attrib[0] & 0xFF; - ypos = (sprline - ypos) & 0xFF; - if (ypos >= (u32)boundheight) - continue; - - s32 xpos = (s32)(attrib[1] << 23) >> 23; - if (xpos <= -boundwidth) - continue; - - u32 rotparamgroup = (attrib[1] >> 9) & 0x1F; - - DoDrawSprite(Rotscale, sprnum, boundwidth, boundheight, width, height, xpos, ypos); - - NumSprites++; - } - else - { - if (attrib[0] & 0x0200) - continue; - - u32 sizeparam = (attrib[0] >> 14) | ((attrib[1] & 0xC000) >> 12); - s32 width = spritewidth[sizeparam]; - s32 height = spriteheight[sizeparam]; - - u32 ypos = attrib[0] & 0xFF; - ypos = (sprline - ypos) & 0xFF; - if (ypos >= (u32)height) - continue; - - s32 xpos = (s32)(attrib[1] << 23) >> 23; - if (xpos <= -width) - continue; - - DoDrawSprite(Normal, sprnum, width, height, xpos, ypos); - - NumSprites++; - } - } - } -} - -template<bool window> -void GPU2D::DrawSprite_Rotscale(u32 num, u32 boundwidth, u32 boundheight, u32 width, u32 height, s32 xpos, s32 ypos) -{ - u16* oam = (u16*)&GPU::OAM[Num ? 0x400 : 0]; - u16* attrib = &oam[num * 4]; - u16* rotparams = &oam[(((attrib[1] >> 9) & 0x1F) * 16) + 3]; - - u32 pixelattr = ((attrib[2] & 0x0C00) << 6) | 0xC0000; - u32 tilenum = attrib[2] & 0x03FF; - u32 spritemode = window ? 0 : ((attrib[0] >> 10) & 0x3); - - u32 ytilefactor; - - s32 centerX = boundwidth >> 1; - s32 centerY = boundheight >> 1; - - if ((attrib[0] & 0x1000) && !window) - { - // apply Y mosaic - pixelattr |= 0x100000; - } - - u32 xoff; - if (xpos >= 0) + if (Num == 0) { - xoff = 0; - if ((xpos+boundwidth) > 256) - boundwidth = 256-xpos; + data = GPU::VRAMFlat_ABG; + mask = 0x7FFFF; } else { - xoff = -xpos; - xpos = 0; - } - - s16 rotA = (s16)rotparams[0]; - s16 rotB = (s16)rotparams[4]; - s16 rotC = (s16)rotparams[8]; - s16 rotD = (s16)rotparams[12]; - - s32 rotX = ((xoff-centerX) * rotA) + ((ypos-centerY) * rotB) + (width << 7); - s32 rotY = ((xoff-centerX) * rotC) + ((ypos-centerY) * rotD) + (height << 7); - - width <<= 8; - height <<= 8; - - u16 color = 0; // transparent in all cases - - if (spritemode == 3) - { - u32 alpha = attrib[2] >> 12; - if (!alpha) return; - alpha++; - - pixelattr |= (0xC0000000 | (alpha << 24)); - - if (DispCnt & 0x40) - { - if (DispCnt & 0x20) - { - // 'reserved' - // draws nothing - - return; - } - else - { - tilenum <<= (7 + ((DispCnt >> 22) & 0x1)); - ytilefactor = ((width >> 8) * 2); - } - } - else - { - if (DispCnt & 0x20) - { - tilenum = ((tilenum & 0x01F) << 4) + ((tilenum & 0x3E0) << 7); - ytilefactor = (256 * 2); - } - else - { - tilenum = ((tilenum & 0x00F) << 4) + ((tilenum & 0x3F0) << 7); - ytilefactor = (128 * 2); - } - } - - u32 pixelsaddr = (Num ? 0x06600000 : 0x06400000) + tilenum; - - for (; xoff < boundwidth;) - { - if ((u32)rotX < width && (u32)rotY < height) - { - color = GPU::ReadVRAM_OBJ<u16>(pixelsaddr + ((rotY >> 8) * ytilefactor) + ((rotX >> 8) << 1)); - - if (color & 0x8000) - { - if (window) OBJWindow[xpos] = 1; - else { OBJLine[xpos] = color | pixelattr; OBJIndex[xpos] = num; } - } - else if (!window) - { - if (OBJLine[xpos] == 0) - { - OBJLine[xpos] = pixelattr & 0x180000; - OBJIndex[xpos] = num; - } - } - } - - rotX += rotA; - rotY += rotC; - xoff++; - xpos++; - } - } - else - { - if (DispCnt & 0x10) - { - tilenum <<= ((DispCnt >> 20) & 0x3); - ytilefactor = (width >> 11) << ((attrib[0] & 0x2000) ? 1:0); - } - else - { - ytilefactor = 0x20; - } - - if (spritemode == 1) pixelattr |= 0x80000000; - else pixelattr |= 0x10000000; - - if (attrib[0] & 0x2000) - { - // 256-color - tilenum <<= 5; - ytilefactor <<= 5; - u32 pixelsaddr = (Num ? 0x06600000 : 0x06400000) + tilenum; - - if (!window) - { - if (!(DispCnt & 0x80000000)) - pixelattr |= 0x1000; - else - pixelattr |= ((attrib[2] & 0xF000) >> 4); - } - - for (; xoff < boundwidth;) - { - if ((u32)rotX < width && (u32)rotY < height) - { - color = GPU::ReadVRAM_OBJ<u8>(pixelsaddr + ((rotY>>11)*ytilefactor) + ((rotY&0x700)>>5) + ((rotX>>11)*64) + ((rotX&0x700)>>8)); - - if (color) - { - if (window) OBJWindow[xpos] = 1; - else { OBJLine[xpos] = color | pixelattr; OBJIndex[xpos] = num; } - } - else if (!window) - { - if (OBJLine[xpos] == 0) - { - OBJLine[xpos] = pixelattr & 0x180000; - OBJIndex[xpos] = num; - } - } - } - - rotX += rotA; - rotY += rotC; - xoff++; - xpos++; - } - } - else - { - // 16-color - tilenum <<= 5; - ytilefactor <<= 5; - u32 pixelsaddr = (Num ? 0x06600000 : 0x06400000) + tilenum; - - if (!window) - { - pixelattr |= 0x1000; - pixelattr |= ((attrib[2] & 0xF000) >> 8); - } - - for (; xoff < boundwidth;) - { - if ((u32)rotX < width && (u32)rotY < height) - { - color = GPU::ReadVRAM_OBJ<u8>(pixelsaddr + ((rotY>>11)*ytilefactor) + ((rotY&0x700)>>6) + ((rotX>>11)*32) + ((rotX&0x700)>>9)); - if (rotX & 0x100) - color >>= 4; - else - color &= 0x0F; - - if (color) - { - if (window) OBJWindow[xpos] = 1; - else { OBJLine[xpos] = color | pixelattr; OBJIndex[xpos] = num; } - } - else if (!window) - { - if (OBJLine[xpos] == 0) - { - OBJLine[xpos] = pixelattr & 0x180000; - OBJIndex[xpos] = num; - } - } - } - - rotX += rotA; - rotY += rotC; - xoff++; - xpos++; - } - } + data = GPU::VRAMFlat_BBG; + mask = 0x1FFFF; } } -template<bool window> -void GPU2D::DrawSprite_Normal(u32 num, u32 width, u32 height, s32 xpos, s32 ypos) +void GPU2D::GetOBJVRAM(u8*& data, u32& mask) { - u16* oam = (u16*)&GPU::OAM[Num ? 0x400 : 0]; - u16* attrib = &oam[num * 4]; - - u32 pixelattr = ((attrib[2] & 0x0C00) << 6) | 0xC0000; - u32 tilenum = attrib[2] & 0x03FF; - u32 spritemode = window ? 0 : ((attrib[0] >> 10) & 0x3); - - u32 wmask = width - 8; // really ((width - 1) & ~0x7) - - if ((attrib[0] & 0x1000) && !window) - { - // apply Y mosaic - pixelattr |= 0x100000; - } - - // yflip - if (attrib[1] & 0x2000) - ypos = height-1 - ypos; - - u32 xoff; - u32 xend = width; - if (xpos >= 0) - { - xoff = 0; - if ((xpos+xend) > 256) - xend = 256-xpos; - } - else - { - xoff = -xpos; - xpos = 0; - } - - u16 color = 0; // transparent in all cases - - if (spritemode == 3) + if (Num == 0) { - // bitmap sprite - - u32 alpha = attrib[2] >> 12; - if (!alpha) return; - alpha++; - - pixelattr |= (0xC0000000 | (alpha << 24)); - - if (DispCnt & 0x40) - { - if (DispCnt & 0x20) - { - // 'reserved' - // draws nothing - - return; - } - else - { - tilenum <<= (7 + ((DispCnt >> 22) & 0x1)); - tilenum += (ypos * width * 2); - } - } - else - { - if (DispCnt & 0x20) - { - tilenum = ((tilenum & 0x01F) << 4) + ((tilenum & 0x3E0) << 7); - tilenum += (ypos * 256 * 2); - } - else - { - tilenum = ((tilenum & 0x00F) << 4) + ((tilenum & 0x3F0) << 7); - tilenum += (ypos * 128 * 2); - } - } - - u32 pixelsaddr = (Num ? 0x06600000 : 0x06400000) + tilenum; - s32 pixelstride; - - if (attrib[1] & 0x1000) // xflip - { - pixelsaddr += (width-1 << 1); - pixelsaddr -= (xoff << 1); - pixelstride = -2; - } - else - { - pixelsaddr += (xoff << 1); - pixelstride = 2; - } - - for (; xoff < xend;) - { - color = GPU::ReadVRAM_OBJ<u16>(pixelsaddr); - - pixelsaddr += pixelstride; - - if (color & 0x8000) - { - if (window) OBJWindow[xpos] = 1; - else { OBJLine[xpos] = color | pixelattr; OBJIndex[xpos] = num; } - } - else if (!window) - { - if (OBJLine[xpos] == 0) - { - OBJLine[xpos] = pixelattr & 0x180000; - OBJIndex[xpos] = num; - } - } - - xoff++; - xpos++; - } + data = GPU::VRAMFlat_AOBJ; + mask = 0x3FFFF; } else { - if (DispCnt & 0x10) - { - tilenum <<= ((DispCnt >> 20) & 0x3); - tilenum += ((ypos >> 3) * (width >> 3)) << ((attrib[0] & 0x2000) ? 1:0); - } - else - { - tilenum += ((ypos >> 3) * 0x20); - } - - if (spritemode == 1) pixelattr |= 0x80000000; - else pixelattr |= 0x10000000; - - if (attrib[0] & 0x2000) - { - // 256-color - tilenum <<= 5; - u32 pixelsaddr = (Num ? 0x06600000 : 0x06400000) + tilenum; - pixelsaddr += ((ypos & 0x7) << 3); - s32 pixelstride; - - if (!window) - { - if (!(DispCnt & 0x80000000)) - pixelattr |= 0x1000; - else - pixelattr |= ((attrib[2] & 0xF000) >> 4); - } - - if (attrib[1] & 0x1000) // xflip - { - pixelsaddr += (((width-1) & wmask) << 3); - pixelsaddr += ((width-1) & 0x7); - pixelsaddr -= ((xoff & wmask) << 3); - pixelsaddr -= (xoff & 0x7); - pixelstride = -1; - } - else - { - pixelsaddr += ((xoff & wmask) << 3); - pixelsaddr += (xoff & 0x7); - pixelstride = 1; - } - - for (; xoff < xend;) - { - color = GPU::ReadVRAM_OBJ<u8>(pixelsaddr); - - pixelsaddr += pixelstride; - - if (color) - { - if (window) OBJWindow[xpos] = 1; - else { OBJLine[xpos] = color | pixelattr; OBJIndex[xpos] = num; } - } - else if (!window) - { - if (OBJLine[xpos] == 0) - { - OBJLine[xpos] = pixelattr & 0x180000; - OBJIndex[xpos] = num; - } - } - - xoff++; - xpos++; - if (!(xoff & 0x7)) pixelsaddr += (56 * pixelstride); - } - } - else - { - // 16-color - tilenum <<= 5; - u32 pixelsaddr = (Num ? 0x06600000 : 0x06400000) + tilenum; - pixelsaddr += ((ypos & 0x7) << 2); - s32 pixelstride; - - if (!window) - { - pixelattr |= 0x1000; - pixelattr |= ((attrib[2] & 0xF000) >> 8); - } - - // TODO: optimize VRAM access!! - // TODO: do xflip better? the 'two pixels per byte' thing makes it a bit shitty - - if (attrib[1] & 0x1000) // xflip - { - pixelsaddr += (((width-1) & wmask) << 2); - pixelsaddr += (((width-1) & 0x7) >> 1); - pixelsaddr -= ((xoff & wmask) << 2); - pixelsaddr -= ((xoff & 0x7) >> 1); - pixelstride = -1; - } - else - { - pixelsaddr += ((xoff & wmask) << 2); - pixelsaddr += ((xoff & 0x7) >> 1); - pixelstride = 1; - } - - for (; xoff < xend;) - { - if (attrib[1] & 0x1000) - { - if (xoff & 0x1) { color = GPU::ReadVRAM_OBJ<u8>(pixelsaddr) & 0x0F; pixelsaddr--; } - else color = GPU::ReadVRAM_OBJ<u8>(pixelsaddr) >> 4; - } - else - { - if (xoff & 0x1) { color = GPU::ReadVRAM_OBJ<u8>(pixelsaddr) >> 4; pixelsaddr++; } - else color = GPU::ReadVRAM_OBJ<u8>(pixelsaddr) & 0x0F; - } - - if (color) - { - if (window) OBJWindow[xpos] = 1; - else { OBJLine[xpos] = color | pixelattr; OBJIndex[xpos] = num; } - } - else if (!window) - { - if (OBJLine[xpos] == 0) - { - OBJLine[xpos] = pixelattr & 0x180000; - OBJIndex[xpos] = num; - } - } - - xoff++; - xpos++; - if (!(xoff & 0x7)) pixelsaddr += ((attrib[1] & 0x1000) ? -28 : 28); - } - } + data = GPU::VRAMFlat_BOBJ; + mask = 0x1FFFF; } } |