common.vsprops

@@ -25,4 +25,8 @@
 		SubSystem="2"
 		RandomizedBaseAddress="1"
 	/>
+	<Tool
+		Name="VCPreBuildEventTool"
+		CommandLine="SubWCRev.exe $(ProjectDir) $(SolutionDir)svnrev_template.h $(ProjectDir)svnrev.h"
+	/>
 </VisualStudioPropertySheet>

gsdx/GS.h

@@ -989,7 +989,7 @@ REG128_SET(GIFPackedReg)
 	GIFPackedNOP	NOP;
 REG_SET_END
 
-struct GIFPath
+__declspec(align(16)) struct GIFPath
 {
 	GIFTag tag; 
 	UINT32 nreg;

gsdx/GSClut.cpp

@@ -25,15 +25,15 @@
 
 GSClut::GSClut()
 {
-	BYTE* p = (BYTE*)_aligned_malloc(256 * 2 * sizeof(WORD) * 2 + 256 * sizeof(DWORD) + 256 * sizeof(UINT64), 16);
+	BYTE* p = (BYTE*)_aligned_malloc(8192, 16);
 
-	m_clut = (WORD*)&p[0];
-	m_buff32 = (DWORD*)&p[1024];
-	m_buff64 = (UINT64*)&p[2048 + 1024];
+	m_clut = (WORD*)&p[0]; // 1k + 1k for buffer overruns (sfex: PSM == PSM_PSMT8, CPSM == PSM_PSMCT32, CSA != 0)
+	m_buff32 = (DWORD*)&p[2048]; // 1k
+	m_buff64 = (UINT64*)&p[4096]; // 2k
 	m_write.dirty = true;
 	m_read.dirty = true;
 
-	for(int i = 0; i < 64; i++)
+	for(int i = 0; i < 16; i++)
 	{
 		for(int j = 0; j < 64; j++)
 		{
@@ -110,8 +110,8 @@ bool GSClut::IsWriting(const GIFRegTEX0& TEX0, const GIFRegTEXCLUT& TEXCLUT)
 	case 3: break;
 	case 4: if(m_CBP[0] == TEX0.CBP) return false; break;
 	case 5: if(m_CBP[1] == TEX0.CBP) return false; break;
-	case 6: return false;
-	case 7: return false;
+	case 6: ASSERT(0); return false;
+	case 7: ASSERT(0); return false;
 	default: __assume(0);
 	}
 
@@ -126,10 +126,10 @@ bool GSClut::Write(const GIFRegTEX0& TEX0, const GIFRegTEXCLUT& TEXCLUT, const G
 	case 1: break;
 	case 2: m_CBP[0] = TEX0.CBP; break;
 	case 3: m_CBP[1] = TEX0.CBP; break;
-	case 4: if(m_CBP[0] == TEX0.CBP) return false;
-	case 5: if(m_CBP[1] == TEX0.CBP) return false;
-	case 6: return false;
-	case 7: return false;
+	case 4: if(m_CBP[0] == TEX0.CBP) return false; break;
+	case 5: if(m_CBP[1] == TEX0.CBP) return false; break;
+	case 6: ASSERT(0); return false;
+	case 7: ASSERT(0); return false;
 	default: __assume(0);
 	}
 
@@ -150,21 +150,29 @@ bool GSClut::Write(const GIFRegTEX0& TEX0, const GIFRegTEXCLUT& TEXCLUT, const G
 
 void GSClut::WriteCLUT32_I8_CSM1(const GIFRegTEX0& TEX0, const GIFRegTEXCLUT& TEXCLUT, const GSLocalMemory* mem)
 {
+	ASSERT(TEX0.CSA == 0);
+
 	WriteCLUT_T32_I8_CSM1(&mem->m_vm32[mem->BlockAddress32(0, 0, TEX0.CBP, 1)], m_clut + (TEX0.CSA << 4));
 }
 
 void GSClut::WriteCLUT32_I4_CSM1(const GIFRegTEX0& TEX0, const GIFRegTEXCLUT& TEXCLUT, const GSLocalMemory* mem)
 {
+	ASSERT(TEX0.CSA < 16);
+
 	WriteCLUT_T32_I4_CSM1(&mem->m_vm32[mem->BlockAddress32(0, 0, TEX0.CBP, 1)], m_clut + (TEX0.CSA << 4));
 }
 
 void GSClut::WriteCLUT16_I8_CSM1(const GIFRegTEX0& TEX0, const GIFRegTEXCLUT& TEXCLUT, const GSLocalMemory* mem)
 {
+	ASSERT(TEX0.CSA < 16);
+
 	WriteCLUT_T16_I8_CSM1(&mem->m_vm16[mem->BlockAddress16(0, 0, TEX0.CBP, 1)], m_clut + (TEX0.CSA << 4));
 }
 
 void GSClut::WriteCLUT16_I4_CSM1(const GIFRegTEX0& TEX0, const GIFRegTEXCLUT& TEXCLUT, const GSLocalMemory* mem)
 {
+	ASSERT(TEX0.CSA < 32);
+
 	WriteCLUT_T16_I4_CSM1(&mem->m_vm16[mem->BlockAddress16(0, 0, TEX0.CBP, 1)], m_clut + (TEX0.CSA << 4));
 }
 

gsdx/GSClut.h

@@ -50,7 +50,7 @@ class GSClut
 
 	typedef void (GSClut::*writeCLUT)(const GIFRegTEX0& TEX0, const GIFRegTEXCLUT& TEXCLUT, const GSLocalMemory* mem);
 
-	writeCLUT m_wc[2][64][64];
+	writeCLUT m_wc[2][16][64];
 
 	void WriteCLUT32_I8_CSM1(const GIFRegTEX0& TEX0, const GIFRegTEXCLUT& TEXCLUT, const GSLocalMemory* mem);
 	void WriteCLUT32_I4_CSM1(const GIFRegTEX0& TEX0, const GIFRegTEXCLUT& TEXCLUT, const GSLocalMemory* mem);

gsdx/GSCrc.cpp

@@ -59,6 +59,7 @@ CRC::Game CRC::m_games[] =
 	{0x086273D2, MetalGearSolid3, FR, false},
 	{0x26A6E286, MetalGearSolid3, EU, false},
 	{0xAA31B5BF, MetalGearSolid3, Unknown, false},
+	{0x9F185CE1, MetalGearSolid3, Unknown, false},
 	{0x278722BF, DBZBT2, US, false},
 	{0xFE961D28, DBZBT2, US, false},
 	{0X0393B6BE, DBZBT2, EU, false},
@@ -97,6 +98,7 @@ CRC::Game CRC::m_games[] =
 	{0xf0a6d880, HarvestMoon, US, true},
 	{0x75c01a04, NamcoXCapcom, US, false},
 	{0xBF6F101F, GiTS, US, false},
+	{0x6BF11378, Onimusha3, US, false},	
 };
 
 CAtlMap<DWORD, CRC::Game*> CRC::m_map;

gsdx/GSCrc.h

@@ -59,7 +59,9 @@ public:
 		JackieChanAdv,
 		HarvestMoon,
 		NamcoXCapcom,
-		GiTS
+		GiTS,
+		Onimusha3,
+		TitleCount
 	};
 
 	enum Region 

gsdx/GSDump.cpp

@@ -38,6 +38,7 @@ GSDump::~GSDump()
 void GSDump::Open(LPCTSTR fn, DWORD crc, const freezeData& fd, const void* regs)
 {
 	m_fp = _tfopen(fn, _T("wb"));
+	m_vsyncs = 0;
 
 	if(m_fp)
 	{
@@ -78,7 +79,7 @@ void GSDump::VSync(int field, bool last, const void* regs)
 		fputc(1, m_fp);
 		fputc(field, m_fp);
 
-		if(!field && last)
+		if((++m_vsyncs & 1) == 0 && last)
 		{
 			fclose(m_fp);
 			m_fp = NULL;

gsdx/GSDump.h

@@ -45,6 +45,7 @@ Regs data (id == 3)
 class GSDump
 {
 	FILE* m_fp;
+	int m_vsyncs;
 
 public:
 	GSDump();

gsdx/GSLocalMemory.h

@@ -870,16 +870,13 @@ public:
 		}
 	}
 
-	__forceinline GSVector4i ReadFrameX(int PSM, const GSVector4i& addr) const
+	__forceinline GSVector4i ReadFrameX(int psm, const GSVector4i& addr) const
 	{
 		GSVector4i c, r, g, b, a;
 
-		// TODO
-
-		switch(PSM)
+		switch(psm)
 		{
-		case PSM_PSMCT32: 
-		case PSM_PSMZ32: 
+		case 0:
 			#if _M_SSE >= 0x401
 			c = addr.gather32_32(m_vm32);
 			#else
@@ -890,8 +887,7 @@ public:
 				(int)ReadPixel32(addr.u32[3]));
 			#endif
 			break;
-		case PSM_PSMCT24: 
-		case PSM_PSMZ24: 
+		case 1:
 			#if _M_SSE >= 0x401
 			c = addr.gather32_32(m_vm32);
 			#else
@@ -903,10 +899,7 @@ public:
 			#endif
 			c = (c & 0x00ffffff) | 0x80000000;
 			break;
-		case PSM_PSMCT16: 
-		case PSM_PSMCT16S: 
-		case PSM_PSMZ16: 
-		case PSM_PSMZ16S: 
+		case 2:
 			#if _M_SSE >= 0x401
 			c = addr.gather32_32(m_vm16);
 			#else
@@ -926,13 +919,13 @@ public:
 		return c;
 	}
 
-	__forceinline GSVector4i ReadZBufX(int PSM, const GSVector4i& addr) const
+	__forceinline GSVector4i ReadZBufX(int psm, const GSVector4i& addr) const
 	{
 		GSVector4i z;
 
-		switch(PSM)
+		switch(psm)
 		{
-		case PSM_PSMZ32: 
+		case 0: 
 			#if _M_SSE >= 0x401
 			z = addr.gather32_32(m_vm32);
 			#else
@@ -943,7 +936,7 @@ public:
 				(int)ReadPixel32(addr.u32[3]));
 			#endif
 			break;
-		case PSM_PSMZ24: 
+		case 1: 
 			#if _M_SSE >= 0x401
 			z = addr.gather32_32(m_vm32) & 0x00ffffff;
 			#else
@@ -955,8 +948,7 @@ public:
 			z = z & 0x00ffffff;
 			#endif
 			break;
-		case PSM_PSMZ16: 
-		case PSM_PSMZ16S: 
+		case 2: 
 			#if _M_SSE >= 0x401
 			z = addr.gather32_32(m_vm16);
 			#else
@@ -975,28 +967,23 @@ public:
 		return z;
 	}
 
-	__forceinline void WriteFrameX(int PSM, const GSVector4i& addr, const GSVector4i& c, const GSVector4i& mask, int pixels)
+	__forceinline void WriteFrameX(int psm, const GSVector4i& addr, const GSVector4i& c, const GSVector4i& mask, int pixels)
 	{
 		GSVector4i rb, ga, tmp;
 
-		switch(PSM)
+		switch(psm)
 		{
-		case PSM_PSMCT32: 
-		case PSM_PSMZ32: 
+		case 0:
 			for(int i = 0; i < pixels; i++)
 				if(mask.u32[i] != 0xffffffff)
 					WritePixel32(addr.u32[i], c.u32[i]); 
 			break; 
-		case PSM_PSMCT24: 
-		case PSM_PSMZ24: 
+		case 1:
 			for(int i = 0; i < pixels; i++)
 				if(mask.u32[i] != 0xffffffff)
 					WritePixel24(addr.u32[i], c.u32[i]); 
 			break; 
-		case PSM_PSMCT16: 
-		case PSM_PSMCT16S: 
-		case PSM_PSMZ16: 
-		case PSM_PSMZ16S: 
+		case 2:
 			rb = c & 0x00f800f8;
 			ga = c & 0x8000f800;
 			tmp = (ga >> 16) | (rb >> 9) | (ga >> 6) | (rb >> 3);
@@ -1010,22 +997,21 @@ public:
 		}
 	}
 
-	__forceinline void WriteZBufX(int PSM, const GSVector4i& addr, const GSVector4i& z, const GSVector4i& mask, int pixels)
+	__forceinline void WriteZBufX(int psm, const GSVector4i& addr, const GSVector4i& z, const GSVector4i& mask, int pixels)
 	{
-		switch(PSM)
+		switch(psm)
 		{
-		case PSM_PSMZ32: 
+		case 0:
 			for(int i = 0; i < pixels; i++)
 				if(mask.u32[i] != 0xffffffff)
 					WritePixel32(addr.u32[i], z.u32[i]); 
 			break; 
-		case PSM_PSMZ24: 
+		case 1:
 			for(int i = 0; i < pixels; i++)
 				if(mask.u32[i] != 0xffffffff)
 					WritePixel24(addr.u32[i], z.u32[i]); 
 			break; 
-		case PSM_PSMZ16: 
-		case PSM_PSMZ16S: 
+		case 2:
 			for(int i = 0; i < pixels; i++)
 				if(mask.u32[i] != 0xffffffff)
 					WritePixel16(addr.u32[i], z.u32[i]); 

gsdx/GSRasterizer.cpp

@@ -60,17 +60,11 @@ GSRasterizer::GSRasterizer(GSState* state, int id, int threads)
 		InitDS_ZPSM(iFPSM, 0) \
 		InitDS_ZPSM(iFPSM, 1) \
 		InitDS_ZPSM(iFPSM, 2) \
-		InitDS_ZPSM(iFPSM, 3) \
 
 	#define InitDS() \
 		InitDS_FPSM(0) \
 		InitDS_FPSM(1) \
 		InitDS_FPSM(2) \
-		InitDS_FPSM(3) \
-		InitDS_FPSM(4) \
-		InitDS_FPSM(5) \
-		InitDS_FPSM(6) \
-		InitDS_FPSM(7) \
 
 	InitDS();
 
@@ -125,8 +119,8 @@ int GSRasterizer::Draw(Vertex* vertices, int count)
 
 	m_sel.dw = 0;
 
-	m_sel.fpsm = GSUtil::EncodeFPSM(context->FRAME.PSM);
-	m_sel.zpsm = GSUtil::EncodeZPSM(context->ZBUF.PSM);
+	m_sel.fpsm = GSUtil::EncodePSM(context->FRAME.PSM);
+	m_sel.zpsm = GSUtil::EncodePSM(context->ZBUF.PSM);
 	m_sel.ztst = context->TEST.ZTE && context->TEST.ZTST > 1 ? context->TEST.ZTST : context->ZBUF.ZMSK ? 0 : 1;
 	m_sel.iip = PRIM->PRIM == GS_POINTLIST || PRIM->PRIM == GS_SPRITE ? 0 : PRIM->IIP;
 	m_sel.tfx = PRIM->TME ? context->TEX0.TFX : 4;
@@ -154,22 +148,19 @@ int GSRasterizer::Draw(Vertex* vertices, int count)
 	}
 
 	m_sel.atst = context->TEST.ATE ? context->TEST.ATST : 1;
-	m_sel.afail = context->TEST.AFAIL;
+	m_sel.afail = context->TEST.ATE ? context->TEST.AFAIL : 0;
 	m_sel.fge = PRIM->FGE;
-	m_sel.rfb = 
-		PRIM->ABE || env.PABE.PABE || context->TEST.DATE ||
-		context->FRAME.FBMSK != 0 && context->FRAME.FBMSK != 0xffffffff || 
-		context->TEST.ATE && context->TEST.ATST != 1 && context->TEST.AFAIL == 3;
 	m_sel.date = context->FRAME.PSM != PSM_PSMCT24 ? context->TEST.DATE : 0;
 	m_sel.abe = env.PABE.PABE ? 2 : PRIM->ABE ? 1 : 0;
 	m_sel.abea = m_sel.abe ? context->ALPHA.A : 0;
 	m_sel.abeb = m_sel.abe ? context->ALPHA.B : 0;
 	m_sel.abec = m_sel.abe ? context->ALPHA.C : 0;
 	m_sel.abed = m_sel.abe ? context->ALPHA.D : 0;
+	m_sel.rfb = m_sel.date || m_sel.abe || m_sel.atst != 1 && m_sel.afail == 3 || context->FRAME.FBMSK != 0 && context->FRAME.FBMSK != 0xffffffff;
 
 	m_dsf = m_ds[m_sel.fpsm][m_sel.zpsm][m_sel.ztst][m_sel.iip];
 
-	CAtlMap<DWORD, DrawScanlinePtr>::CPair* pair = m_dsmap2.Lookup(m_sel);
+	CRBMap<DWORD, DrawScanlinePtr>::CPair* pair = m_dsmap2.Lookup(m_sel);
 
 	if(pair)
 	{
@@ -183,7 +174,7 @@ int GSRasterizer::Draw(Vertex* vertices, int count)
 		{
 			m_dsf = pair->m_value;
 
-			m_dsmap2[pair->m_key] = pair->m_value;
+			m_dsmap2.SetAt(pair->m_key, pair->m_value);
 		}
 		else if(!pair)
 		{
@@ -193,11 +184,11 @@ int GSRasterizer::Draw(Vertex* vertices, int count)
 				m_sel.tfx, m_sel.tcc, m_sel.fst, m_sel.ltf, 
 				m_sel.atst, m_sel.afail, m_sel.fge, m_sel.rfb, m_sel.date, m_sel.abe);
 
-			m_dsmap[m_sel] = NULL;
+			m_dsmap.SetAt(m_sel, NULL);
 
 			if(FILE* fp = _tfopen(_T("c:\\1.txt"), _T("w")))
 			{
-				POSITION pos = m_dsmap.GetStartPosition();
+				POSITION pos = m_dsmap.GetHeadPosition();
 
 				while(pos) 
 				{
@@ -205,7 +196,7 @@ int GSRasterizer::Draw(Vertex* vertices, int count)
 
 					if(!pair->m_value)
 					{
-						_ftprintf(fp, _T("m_dsmap[0x%08x] = &GSRasterizer::DrawScanlineEx<0x%08x>;\n"), pair->m_key, pair->m_key);
+						_ftprintf(fp, _T("m_dsmap.SetAt(0x%08x, &GSRasterizer::DrawScanlineEx<0x%08x>);\n"), pair->m_key, pair->m_key);
 					}
 				}
 
@@ -828,25 +819,9 @@ void GSRasterizer::SetupScanline(const Vertex& dv)
 	}
 }
 
-template<int iFPSM, int iZPSM, int ztst, int iip>
+template<DWORD fpsm, DWORD zpsm, DWORD ztst, DWORD iip>
 void GSRasterizer::DrawScanline(int top, int left, int right, const Vertex& v)	
 {
-/*
-extern UINT64 g_slp1;
-extern UINT64 g_slp2;
-extern UINT64 g_slp3;
-extern UINT64 g_slp4;
-{
-int steps = right - left;
-for(; steps >= 4; steps -= 4) g_slp4++;
-if(steps == 1) g_slp1++;
-else if(steps == 2) g_slp2++;
-else if(steps == 3) g_slp3++;
-}
-*/
-	int fpsm = GSUtil::DecodeFPSM(iFPSM);
-	int zpsm = GSUtil::DecodeZPSM(iZPSM);
-
 	GSVector4i fa_base = m_slenv.fbco[top];
 	GSVector4i* fa_offset = (GSVector4i*)&m_slenv.fo[left];
 
@@ -887,18 +862,24 @@ else if(steps == 3) g_slp3++;
 		GSVector4i test = GSVector4i::zero();
 
 		GSVector4i zs = (GSVector4i(z * 0.5f) << 1) | (GSVector4i(z) & GSVector4i::one());
-		GSVector4i zd;
 
 		if(ztst > 1)
 		{
-			zd = m_state->m_mem.ReadZBufX(zpsm, za);
+			GSVector4i zd = m_state->m_mem.ReadZBufX(zpsm, za);
 
-			GSVector4i offset = GSVector4i::x80000000();
+			GSVector4i zso = zs;
+			GSVector4i zdo = zd;
+
+			if(zpsm == 0)
+			{
+				zso = zs - GSVector4i::x80000000();
+				zdo = zd - GSVector4i::x80000000();
+			}
 
 			switch(ztst)
 			{
-			case 2: test = (zs - offset) < (zd - offset); break; // ge
-			case 3: test = (zs - offset) <= (zd - offset); break; // g
+			case 2: test = zso < zdo; break; // ge
+			case 3: test = zso <= zdo; break; // g
 			default: __assume(0);
 			}
 
@@ -1099,10 +1080,10 @@ else if(steps == 3) g_slp3++;
 			c[10] = GSVector4::zero();
 			c[11] = m_slenv.afix;
 
-			int abea = m_sel.abea;
-			int abeb = m_sel.abeb;
-			int abec = m_sel.abec;
-			int abed = m_sel.abed;
+			DWORD abea = m_sel.abea;
+			DWORD abeb = m_sel.abeb;
+			DWORD abec = m_sel.abec;
+			DWORD abed = m_sel.abed;
 
 			GSVector4 r = (c[abea*4 + 0] - c[abeb*4 + 0]).mod2x(c[abec*4 + 3]) + c[abed*4 + 0];
 			GSVector4 g = (c[abea*4 + 1] - c[abeb*4 + 1]).mod2x(c[abec*4 + 3]) + c[abed*4 + 1];

gsdx/GSRasterizer.h

@@ -86,29 +86,29 @@ private:
 	{
 		struct
 		{
-			DWORD fpsm:3; // 0
-			DWORD zpsm:2; // 3
-			DWORD ztst:2; // 5 (0: off, 1: write, 2: write + test (ge), 3: write + test (g))
-			DWORD iip:1; // 7
-			DWORD tfx:3; // 8
-			DWORD tcc:1; // 11
-			DWORD fst:1; // 12
-			DWORD ltf:1; // 13
-			DWORD atst:3; // 14
-			DWORD afail:2; // 17
-			DWORD fge:1; // 19
-			DWORD rfb:1; // 20
-			DWORD date:1; // 21
-			DWORD abe:2; // 22
-			DWORD abea:2; // 24
-			DWORD abeb:2; // 26
-			DWORD abec:2; // 28
-			DWORD abed:2; // 30
+			DWORD fpsm:2; // 0
+			DWORD zpsm:2; // 2
+			DWORD ztst:2; // 4 (0: off, 1: write, 2: write + test (ge), 3: write + test (g))
+			DWORD iip:1; // 6
+			DWORD tfx:3; // 7
+			DWORD tcc:1; // 10
+			DWORD fst:1; // 11
+			DWORD ltf:1; // 12
+			DWORD atst:3; // 13
+			DWORD afail:2; // 16
+			DWORD fge:1; // 18
+			DWORD date:1; // 19
+			DWORD abe:2; // 20
+			DWORD abea:2; // 22
+			DWORD abeb:2; // 24
+			DWORD abec:2; // 26
+			DWORD abed:2; // 28
+			DWORD rfb:1; // 30
 		};
 
 		DWORD dw;
 
-		operator DWORD() {return dw & 0xffffffff;}
+		operator DWORD() {return dw & 0x7fffffff;}
 	};
 
 	CRect m_scissor;
@@ -137,10 +137,10 @@ private:
 
 	typedef void (GSRasterizer::*DrawScanlinePtr)(int top, int left, int right, const Vertex& v);
 
-	DrawScanlinePtr m_ds[8][4][4][2], m_dsf;
-	CAtlMap<DWORD, DrawScanlinePtr> m_dsmap, m_dsmap2;
+	DrawScanlinePtr m_ds[4][4][4][2], m_dsf;
+	CRBMap<DWORD, DrawScanlinePtr> m_dsmap, m_dsmap2;
 
-	template<int iFPSM, int iZPSM, int iZTST, int iIIP>
+	template<DWORD fpsm, DWORD zpsm, DWORD ztst, DWORD iip>
 	void DrawScanline(int top, int left, int right, const Vertex& v);
 
 	void InitEx();

gsdx/GSRenderer.h

@@ -441,7 +441,7 @@ public:
 
 		fn.Format(_T("%s_%s"), path, CTime::GetCurrentTime().Format(_T("%Y%m%d%H%M%S")));
 
-		if((::GetAsyncKeyState(VK_SHIFT) & 0x8000) && !m_dump && m_field == 0)
+		if((::GetAsyncKeyState(VK_SHIFT) & 0x8000) && !m_dump)
 		{
 			freezeData fd;
 			fd.size = 0;

gsdx/GSRendererHW10.cpp

@@ -34,6 +34,19 @@ GSRendererHW10::GSRendererHW10(BYTE* base, bool mt, void (*irq)(), int nloophack
 	{
 		m_fpDrawingKickHandlers[i] = (DrawingKickHandler)&GSRendererHW10::DrawingKick;
 	}
+
+	m_fpDrawingKickHandlers[GS_POINTLIST] = (DrawingKickHandler)&GSRendererHW10::DrawingKickPoint;
+
+	#if _M_SSE >= 0x401
+
+	m_fpDrawingKickHandlers[GS_LINELIST] = (DrawingKickHandler)&GSRendererHW10::DrawingKickLine;
+	m_fpDrawingKickHandlers[GS_LINESTRIP] = (DrawingKickHandler)&GSRendererHW10::DrawingKickLine;
+	m_fpDrawingKickHandlers[GS_TRIANGLELIST] = (DrawingKickHandler)&GSRendererHW10::DrawingKickTriangle;
+	m_fpDrawingKickHandlers[GS_TRIANGLESTRIP] = (DrawingKickHandler)&GSRendererHW10::DrawingKickTriangle;
+	m_fpDrawingKickHandlers[GS_TRIANGLEFAN] = (DrawingKickHandler)&GSRendererHW10::DrawingKickTriangle;
+	m_fpDrawingKickHandlers[GS_SPRITE] = (DrawingKickHandler)&GSRendererHW10::DrawingKickSprite;
+
+	#endif
 }
 
 bool GSRendererHW10::Create(LPCTSTR title)
@@ -120,10 +133,82 @@ void GSRendererHW10::VertexKick(bool skip)
 	__super::VertexKick(skip);
 }
 
+int GSRendererHW10::ScissorTest(const GSVector4i& p0, const GSVector4i& p1)
+{
+	GSVector4i scissor = m_context->scissor.dx10;
+
+	GSVector4i v0 = p0 < scissor;
+	GSVector4i v1 = p1 > scissor.zwxy();
+
+	return (v0 | v1).mask() & 0xff;
+}
+
+void GSRendererHW10::DrawingKickPoint(GSVertexHW10* v, int& count)
+{
+	GSVector4i v0 = GSVector4i::load((int)v[0].p.xy).upl16();
+
+	GSVector4i p0 = v0;
+	GSVector4i p1 = v0;
+
+	if(ScissorTest(p0, p1))
+	{
+		count = 0;
+		return;
+	}
+}
+
+#if _M_SSE >= 0x401
+
+void GSRendererHW10::DrawingKickLine(GSVertexHW10* v, int& count)
+{
+	GSVector4i v0 = GSVector4i::load((int)v[0].p.xy);
+	GSVector4i v1 = GSVector4i::load((int)v[1].p.xy);
+
+	GSVector4i p0 = v0.max_u16(v1).upl16();
+	GSVector4i p1 = v0.min_u16(v1).upl16();
+
+	if(ScissorTest(p0, p1))
+	{
+		count = 0;
+		return;
+	}
+}
+
+void GSRendererHW10::DrawingKickTriangle(GSVertexHW10* v, int& count)
+{
+	GSVector4i v0 = GSVector4i::load((int)v[0].p.xy);
+	GSVector4i v1 = GSVector4i::load((int)v[1].p.xy);
+	GSVector4i v2 = GSVector4i::load((int)v[2].p.xy);
+
+	GSVector4i p0 = v0.max_u16(v1).max_u16(v2).upl16();
+	GSVector4i p1 = v0.min_u16(v1).min_u16(v2).upl16();
+
+	if(ScissorTest(p0, p1))
+	{
+		count = 0;
+		return;
+	}
+}
+
+void GSRendererHW10::DrawingKickSprite(GSVertexHW10* v, int& count)
+{
+	GSVector4i v0 = GSVector4i::load((int)v[0].p.xy);
+	GSVector4i v1 = GSVector4i::load((int)v[1].p.xy);
+
+	GSVector4i p0 = v0.max_u16(v1).upl16();
+	GSVector4i p1 = v0.min_u16(v1).upl16();
+
+	if(ScissorTest(p0, p1))
+	{
+		count = 0;
+		return;
+	}
+}
+
+#endif
+
 void GSRendererHW10::DrawingKick(GSVertexHW10* v, int& count)
 {
-	// TODO
-
 	GSVector4i scissor = m_context->scissor.dx10;
 
 	switch(count)

gsdx/GSRendererHW10.h

@@ -35,6 +35,18 @@ protected:
 	void Draw();
 	bool WrapZ(DWORD maxz);
 
+	__forceinline int ScissorTest(const GSVector4i& p0, const GSVector4i& p1);
+
+	void DrawingKickPoint(GSVertexHW10* v, int& count);
+
+	#if _M_SSE >= 0x401
+
+	void DrawingKickLine(GSVertexHW10* v, int& count);
+	void DrawingKickTriangle(GSVertexHW10* v, int& count);
+	void DrawingKickSprite(GSVertexHW10* v, int& count);
+
+	#endif
+
 	struct
 	{
 		CComPtr<ID3D10DepthStencilState> dss;

gsdx/GSState.cpp

@@ -115,8 +115,6 @@ GSState::GSState(BYTE* base, bool mt, void (*irq)(), int nloophack)
 	m_maxbytes = 1024 * 1024 * 4;
 	m_buff = (BYTE*)_aligned_malloc(m_maxbytes, 16);
 
-	m_path = (GIFPath*)_aligned_malloc(sizeof(m_path[0]) * 3, 16);
-
 	Reset();
 
 	ResetHandlers();
@@ -125,7 +123,6 @@ GSState::GSState(BYTE* base, bool mt, void (*irq)(), int nloophack)
 GSState::~GSState()
 {
 	_aligned_free(m_buff);
-	_aligned_free(m_path);
 }
 
 void GSState::Reset()
@@ -597,23 +594,19 @@ template<int i> void GSState::GIFRegHandlerTEX0(GIFReg* r)
 
 	m_env.CTXT[i].TEX0 = r->TEX0;
 
-	// ASSERT(m_env.CTXT[i].TEX0.TW <= 10 && m_env.CTXT[i].TEX0.TH <= 10 && (m_env.CTXT[i].TEX0.CPSM & ~0xa) == 0);
-
 	if(m_env.CTXT[i].TEX0.TW > 10) m_env.CTXT[i].TEX0.TW = 10;
 	if(m_env.CTXT[i].TEX0.TH > 10) m_env.CTXT[i].TEX0.TH = 10;
 
 	m_env.CTXT[i].TEX0.CPSM &= 0xa; // 1010b
 
-	m_env.CTXT[i].ttbl = &GSLocalMemory::m_psm[m_env.CTXT[i].TEX0.PSM];
-
-	FlushWrite();
-
-	m_mem.m_clut.Write(r->TEX0, m_env.TEXCLUT, &m_mem);
-
 	if((m_env.CTXT[i].TEX0.TBW & 1) && (m_env.CTXT[i].TEX0.PSM == PSM_PSMT8 || m_env.CTXT[i].TEX0.PSM == PSM_PSMT4))
 	{
-		m_env.CTXT[i].TEX0.TBW &= ~1;
+		m_env.CTXT[i].TEX0.TBW &= ~1; // GS User 2.6
 	}
+
+	m_env.CTXT[i].ttbl = &GSLocalMemory::m_psm[m_env.CTXT[i].TEX0.PSM];
+
+	m_mem.m_clut.Write(m_env.CTXT[i].TEX0, m_env.TEXCLUT, &m_mem);
 }
 
 template<int i> void GSState::GIFRegHandlerCLAMP(GIFReg* r)
@@ -1257,7 +1250,7 @@ template<int index> void GSState::Transfer(BYTE* mem, UINT32 size)
 
 			if(path.tag.PRE)
 			{
-				ASSERT(path.tag.FLG != GIF_FLG_IMAGE); // kingdom hearts
+				ASSERT(path.tag.FLG != GIF_FLG_IMAGE); // kingdom hearts, ffxii
 
 				if((path.tag.FLG & 2) == 0)
 				{
@@ -2092,6 +2085,24 @@ bool GSC_GiTS(const GSFrameInfo& fi, int& skip)
 	return true;
 }
 
+bool GSC_Onimusha3(const GSFrameInfo& fi, int& skip)
+{
+	if(skip == 0)
+	{
+		if(fi.TME && (fi.FBP == 0x00e00 && fi.TBP0 == 0x00700 /*|| fi.FBP == 0x01000 && fi.TBP0 == 0x00e00*/))
+		{
+//			skip = 1000;
+		}
+	}
+
+	if(fi.TME && (fi.FBP == 0x00700 && fi.TBP0 == 0x00e00 /*|| fi.FBP == 0x00e00 && fi.TBP0 == 0x01000*/))
+	{
+//		skip = 1;
+	}
+
+	return true;
+}
+
 bool GSState::IsBadFrame(int& skip)
 {
 	GSFrameInfo fi;
@@ -2102,10 +2113,15 @@ bool GSState::IsBadFrame(int& skip)
 	fi.TBP0 = m_context->TEX0.TBP0;
 	fi.TPSM = m_context->TEX0.PSM;
 
-	static CAtlMap<CRC::Title, GetSkipCount> map;
+	static GetSkipCount map[CRC::TitleCount];
+	static bool inited = false;
 
-	if(map.IsEmpty())
+	if(!inited)
 	{
+		inited = true;
+
+		memset(map, 0, sizeof(map));
+		
 		map[CRC::Okami] = GSC_Okami;
 		map[CRC::MetalGearSolid3] = GSC_MetalGearSolid3;
 		map[CRC::DBZBT2] = GSC_DBZBT2;
@@ -2129,14 +2145,14 @@ bool GSState::IsBadFrame(int& skip)
 		map[CRC::GodOfWar] = GSC_GodOfWar;
 		map[CRC::GodOfWar2] = GSC_GodOfWar;
 		map[CRC::GiTS] = GSC_GiTS;
+		map[CRC::Onimusha3] = GSC_Onimusha3;
 	}
 
-	if(CAtlMap<CRC::Title, GetSkipCount>::CPair* pair = map.Lookup(m_game.title))
+	GetSkipCount gsc = map[m_game.title];
+
+	if(gsc && !gsc(fi, skip))
 	{
-		if(!pair->m_value(fi, skip))
-		{
-			return false;
-		}
+		return false;
 	}
 
 	if(skip == 0)

gsdx/GSState.h

@@ -141,7 +141,7 @@ public:
 	GSRegBUSDIR*	BUSDIR;
 	GSRegSIGLBLID*	SIGLBLID;
 
-	GIFPath* m_path;
+	GIFPath m_path[3];
 	GSLocalMemory m_mem;
 	GSDrawingEnvironment m_env;
 	GSDrawingContext* m_context;

gsdx/GSTextureCache.h

@@ -107,12 +107,12 @@ public:
 	class GSTexture : public GSSurface
 	{
 	protected:
-		bool GetDirtyRect(CRect& r)
+		bool GetDirtyRect(CRect& rr)
 		{
 			int w = 1 << m_TEX0.TW;
 			int h = 1 << m_TEX0.TH;
 
-			r = CRect(0, 0, w, h);
+			CRect r(0, 0, w, h);
 
 			POSITION pos = m_dirty.GetHeadPosition();
 
@@ -172,7 +172,14 @@ public:
 				r |= m_valid;
 			}
 
-			return !r.IsRectEmpty();
+			if(r.IsRectEmpty())
+			{
+				return false;
+			}
+
+			rr = r;
+
+			return true;
 		}
 
 	public:

gsdx/GSUtil.h

@@ -67,65 +67,24 @@ public:
 	static bool IsRectInRectH(const CRect& inner, const CRect& outer);
 	static bool IsRectInRectV(const CRect& inner, const CRect& outer);
 	
-	static int EncodeFPSM(int psm)
+	static int EncodePSM(int psm)
 	{
 		switch(psm)
 		{
-		case PSM_PSMCT32: return 0;
-		case PSM_PSMCT24: return 1;
-		case PSM_PSMCT16: return 2;
-		case PSM_PSMCT16S: return 3;
-		case PSM_PSMZ32: return 4;
-		case PSM_PSMZ24: return 5;
-		case PSM_PSMZ16: return 6;
-		case PSM_PSMZ16S: return 7;
+		case PSM_PSMCT32: 
+		case PSM_PSMZ32: 
+			return 0;
+		case PSM_PSMCT24: 
+		case PSM_PSMZ24:
+			return 1;
+		case PSM_PSMCT16: 
+		case PSM_PSMCT16S: 
+		case PSM_PSMZ16:
+		case PSM_PSMZ16S:
+			return 2;
+		default:
+			return 3;
 		}
-
-		return -1;
 	}
-
-	static int DecodeFPSM(int index)
-	{
-		switch(index)
-		{
-		case 0: return PSM_PSMCT32;
-		case 1: return PSM_PSMCT24;
-		case 2: return PSM_PSMCT16;
-		case 3: return PSM_PSMCT16S;
-		case 4: return PSM_PSMZ32;
-		case 5: return PSM_PSMZ24;
-		case 6: return PSM_PSMZ16;
-		case 7: return PSM_PSMZ16S;
-		}
-
-		return -1;
-	}
-
-	static int EncodeZPSM(int psm)
-	{
-		switch(psm)
-		{
-		case PSM_PSMZ32: return 0;
-		case PSM_PSMZ24: return 1;
-		case PSM_PSMZ16: return 2;
-		case PSM_PSMZ16S: return 3;
-		}
-
-		return -1;
-	}
-
-	static int DecodeZPSM(int index)
-	{
-		switch(index)
-		{
-		case 0: return PSM_PSMZ32;
-		case 1: return PSM_PSMZ24;
-		case 2: return PSM_PSMZ16;
-		case 3: return PSM_PSMZ16S;
-		}
-
-		return -1;
-	}
-
 };
 

gsdx/GSVector.h

@@ -250,6 +250,78 @@ public:
 
 	#endif
 
+	#if _M_SSE >= 0x401
+
+	GSVector4i min_i8(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_min_epi8(m, a));
+	}
+
+	GSVector4i max_i8(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_max_epi8(m, a));
+	}
+
+	#endif
+
+	GSVector4i min_i16(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_min_epi16(m, a));
+	}
+
+	GSVector4i max_i16(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_max_epi16(m, a));
+	}
+
+	#if _M_SSE >= 0x401
+
+	GSVector4i min_i32(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_min_epi32(m, a));
+	}
+
+	GSVector4i max_i32(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_max_epi32(m, a));
+	}
+
+	#endif
+
+	GSVector4i min_u8(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_min_epu8(m, a));
+	}
+
+	GSVector4i max_u8(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_max_epu8(m, a));
+	}
+
+	#if _M_SSE >= 0x401
+
+	GSVector4i min_u16(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_min_epu16(m, a));
+	}
+
+	GSVector4i max_u16(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_max_epu16(m, a));
+	}
+
+	GSVector4i min_u32(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_min_epu32(m, a));
+	}
+
+	GSVector4i max_u32(const GSVector4i& a) const
+	{
+		return GSVector4i(_mm_max_epu32(m, a));
+	}
+
+	#endif
+
 	GSVector4i blend8(const GSVector4i& a, const GSVector4i& mask) const
 	{
 		return GSVector4i(_mm_blendv_epi8(m, a, mask));

gsdx/GSVertexHW.h

@@ -57,7 +57,7 @@ __declspec(align(16)) union GSVertexHW10
 
 		union
 		{
-			struct {WORD x, y; DWORD z;} p;
+			struct {union {struct {WORD x, y;}; DWORD xy;}; DWORD z;} p;
 			GIFRegXYZ XYZ;
 		};
 		

gsdx/GSdx.cpp

@@ -26,6 +26,7 @@
 #include "GSRendererSW.h"
 #include "GSRendererNull.h"
 #include "GSSettingsDlg.h"
+#include "svnrev.h"
 
 //
 //	Note!
@@ -139,7 +140,11 @@ EXPORT_C_(UINT32) PS2EgetLibType()
 
 EXPORT_C_(char*) PS2EgetLibName()
 {
-	CString str = _T("GSdx");
+	CString str;
+
+	str.Format(_T("GSdx %d"), SVN_REV);
+
+	if(SVN_MODS) str += _T("m");
 
 #if _M_AMD64
 	str += _T(" 64-bit");

gsdx/GSdx_vs2008.vcproj

@@ -1781,6 +1781,10 @@
 				RelativePath=".\stdafx.h"
 				>
 			</File>
+			<File
+				RelativePath=".\svnrev.h"
+				>
+			</File>
 		</Filter>
 		<Filter
 			Name="Resource Files"

svnrev_template.h

@@ -0,0 +1,2 @@
+#define SVN_REV $WCREV$ 
+#define SVN_MODS $WCMODS?1:0$

xpad/xpad_vs2008.vcproj

@@ -1202,6 +1202,10 @@
 				RelativePath=".\stdafx.h"
 				>
 			</File>
+			<File
+				RelativePath=".\svnrev.h"
+				>
+			</File>
 			<File
 				RelativePath=".\xpad.h"
 				>