Merge branch 'master' into wii-network

# By Ryan Houdek (185) and others
# Via degasus (12) and others
* master: (625 commits)
  Revert "Don't open/close file for every file operation." as it was crashing PokePark in Windows builds.
  Array overrun fixed in VertexShaderCache for the DX11 plugin.
  Fixed DSPTool build.
  Windows build fix
  Go back to assuming every HID device is a wiimote on Windows. Fixed issue 6117. Unfixed issue 6031.
  VideoSoftware: Improve fog range adjustment by using less magic and more comments.
  revert RasterFont for VideoSoftware
  ogl: fix virtual xfb
  Windows build fix from web interface...
  Adjusted the audio loop criteria, using >= on the Wii and == on GC.  This fixes the audio static that occurred in Wii games after hours of play.
  Forced the exception check only for ARAM DMA transfers. Removed the Eternal Darkness boot hack and replaced it with an exception check.
  VideoSoftware: Implement fog range adjustment, fixing issue 6147.
  implement 4xSSAA for OGL
  move ogl-only settings into backend
  Fix description of disable fog, and move it to enhancements tab.
  Reverted rd76ca5783743 as it was made obsolete by r1d550f4496e4.
  Removed the tracking of the FIFO Writes as it was made obsolete by r1d550f4496e4.
  Forced the external exception check to occur sooner by changing the downcount.
  Mark the Direct3D9 backend deprecated.
  Prefer D3D11 and OpenGL over D3D9 by default.
  ...

Conflicts:
	CMakeLists.txt
	Source/Core/Common/Common.vcxproj.filters
	Source/Core/Common/Src/CommonPaths.h
	Source/Core/Core/Core.vcxproj.filters
	Source/Core/Core/Src/Core.cpp
	Source/Core/Core/Src/IPC_HLE/WII_IPC_HLE_Device_FileIO.cpp
	Source/VSProps/Dolphin.Win32.props
	Source/VSProps/Dolphin.x64.props

1 files changed, 266 insertions, 157 deletions

diff --git a/Source/Core/VideoCommon/Src/PixelShaderGen.cpp b/Source/Core/VideoCommon/Src/PixelShaderGen.cpp
index 448501aad0..d863a6e0a9 100644
--- a/Source/Core/VideoCommon/Src/PixelShaderGen.cpp
+++ b/Source/Core/VideoCommon/Src/PixelShaderGen.cpp
@@ -173,7 +173,7 @@ void GetPixelShaderId(PIXELSHADERUID *uid, DSTALPHA_MODE dstAlphaMode, u32 compo
 		*ptr++ = components;
 	}
 
-	uid->num_values = ptr - uid->values;
+	uid->num_values = int(ptr - uid->values);
 }
 
 void GetSafePixelShaderId(PIXELSHADERUIDSAFE *uid, DSTALPHA_MODE dstAlphaMode, u32 components)
@@ -205,7 +205,7 @@ void GetSafePixelShaderId(PIXELSHADERUIDSAFE *uid, DSTALPHA_MODE dstAlphaMode, u
 
 	*ptr++ = bpmem.tevindref.hex; // 31
 
-	for (unsigned int i = 0; i < bpmem.genMode.numtevstages+1u; ++i) // up to 16 times
+	for (u32 i = 0; i < bpmem.genMode.numtevstages+1u; ++i) // up to 16 times
 	{
 		*ptr++ = bpmem.combiners[i].colorC.hex; // 32+5*i
 		*ptr++ = bpmem.combiners[i].alphaC.hex; // 33+5*i
@@ -252,7 +252,8 @@ void ValidatePixelShaderIDs(API_TYPE api, PIXELSHADERUIDSAFE old_id, const std::
 			static int num_failures = 0;
 			char szTemp[MAX_PATH];
 			sprintf(szTemp, "%spsuid_mismatch_%04i.txt", File::GetUserPath(D_DUMP_IDX).c_str(), num_failures++);
-			std::ofstream file(szTemp);
+			std::ofstream file;
+			OpenFStream(file, szTemp, std::ios_base::out);
 			file << msg;
 			file << "\n\nOld shader code:\n" << old_code;
 			file << "\n\nNew shader code:\n" << new_code;
@@ -275,7 +276,7 @@ void ValidatePixelShaderIDs(API_TYPE api, PIXELSHADERUIDSAFE old_id, const std::
 static void WriteStage(char *&p, int n, API_TYPE ApiType);
 static void SampleTexture(char *&p, const char *destination, const char *texcoords, const char *texswap, int texmap, API_TYPE ApiType);
 // static void WriteAlphaCompare(char *&p, int num, int comp);
-static void WriteAlphaTest(char *&p, API_TYPE ApiType,DSTALPHA_MODE dstAlphaMode);
+static void WriteAlphaTest(char *&p, API_TYPE ApiType,DSTALPHA_MODE dstAlphaMode, bool per_pixel_depth);
 static void WriteFog(char *&p);
 
 static const char *tevKSelTableC[] = // KCSEL
@@ -373,38 +374,38 @@ static const char *tevOpTable[] = { // TEV
 
 static const char *tevCInputTable[] = // CC
 {
-	"(prev.rgb)",               // CPREV,
+	"(prev.rgb)",         // CPREV,
 	"(prev.aaa)",         // APREV,
-	"(c0.rgb)",                 // C0,
+	"(c0.rgb)",           // C0,
 	"(c0.aaa)",           // A0,
-	"(c1.rgb)",                 // C1,
+	"(c1.rgb)",           // C1,
 	"(c1.aaa)",           // A1,
-	"(c2.rgb)",                 // C2,
+	"(c2.rgb)",           // C2,
 	"(c2.aaa)",           // A2,
-	"(textemp.rgb)",            // TEXC,
+	"(textemp.rgb)",      // TEXC,
 	"(textemp.aaa)",      // TEXA,
-	"(rastemp.rgb)",            // RASC,
+	"(rastemp.rgb)",      // RASC,
 	"(rastemp.aaa)",      // RASA,
 	"float3(1.0f, 1.0f, 1.0f)",              // ONE
-	"float3(0.5f, 0.5f, 0.5f)",                 // HALF
-	"(konsttemp.rgb)", //"konsttemp.rgb",        // KONST
+	"float3(0.5f, 0.5f, 0.5f)",              // HALF
+	"(konsttemp.rgb)", //"konsttemp.rgb",    // KONST
 	"float3(0.0f, 0.0f, 0.0f)",              // ZERO
 	///aded extra values to map clamped values
-	"(cprev.rgb)",               // CPREV,
-	"(cprev.aaa)",         // APREV,
-	"(cc0.rgb)",                 // C0,
-	"(cc0.aaa)",           // A0,
-	"(cc1.rgb)",                 // C1,
-	"(cc1.aaa)",           // A1,
-	"(cc2.rgb)",                 // C2,
-	"(cc2.aaa)",           // A2,
-	"(textemp.rgb)",            // TEXC,
+	"(cprev.rgb)",        // CPREV,
+	"(cprev.aaa)",        // APREV,
+	"(cc0.rgb)",          // C0,
+	"(cc0.aaa)",          // A0,
+	"(cc1.rgb)",          // C1,
+	"(cc1.aaa)",          // A1,
+	"(cc2.rgb)",          // C2,
+	"(cc2.aaa)",          // A2,
+	"(textemp.rgb)",      // TEXC,
 	"(textemp.aaa)",      // TEXA,
-	"(crastemp.rgb)",            // RASC,
-	"(crastemp.aaa)",      // RASA,
+	"(crastemp.rgb)",     // RASC,
+	"(crastemp.aaa)",     // RASA,
 	"float3(1.0f, 1.0f, 1.0f)",              // ONE
-	"float3(0.5f, 0.5f, 0.5f)",                 // HALF
-	"(ckonsttemp.rgb)", //"konsttemp.rgb",        // KONST
+	"float3(0.5f, 0.5f, 0.5f)",              // HALF
+	"(ckonsttemp.rgb)", //"konsttemp.rgb",   // KONST
 	"float3(0.0f, 0.0f, 0.0f)",              // ZERO
 	"PADERROR1", "PADERROR2", "PADERROR3", "PADERROR4"
 };
@@ -424,7 +425,7 @@ static const char *tevAInputTable[] = // CA
 	"cc0",              // A0,
 	"cc1",              // A1,
 	"cc2",              // A2,
-	"textemp",         // TEXA,
+	"textemp",          // TEXA,
 	"crastemp",         // RASA,
 	"ckonsttemp",       // KONST,  (hw1 had quarter)
 	"float4(0.0f, 0.0f, 0.0f, 0.0f)", // ZERO
@@ -439,8 +440,8 @@ static const char *tevRasTable[] =
 	"ERROR13", //2
 	"ERROR14", //3
 	"ERROR15", //4
-	"alphabump", // use bump alpha
-	"(alphabump*(255.0f/248.0f))", //normalized
+	"float4(alphabump,alphabump,alphabump,alphabump)", // use bump alpha
+	"(float4(alphabump,alphabump,alphabump,alphabump)*(255.0f/248.0f))", //normalized
 	"float4(0.0f, 0.0f, 0.0f, 0.0f)", // zero
 };
 
@@ -484,6 +485,24 @@ static void BuildSwapModeTable()
 	}
 }
 
+const char* WriteRegister(API_TYPE ApiType, const char *prefix, const u32 num)
+{
+	if (ApiType == API_OPENGL)
+		return ""; // Nothing to do here
+	static char result[64];
+	sprintf(result, " : register(%s%d)", prefix, num);
+	return result;
+}
+
+const char *WriteLocation(API_TYPE ApiType)
+{
+	if (g_ActiveConfig.backend_info.bSupportsGLSLUBO)
+		return "";
+	static char result[64];
+	sprintf(result, "uniform ");
+	return result;
+}
+
 const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType, u32 components)
 {
 	setlocale(LC_NUMERIC, "C"); // Reset locale for compilation
@@ -493,6 +512,8 @@ const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType
 	int numStages = bpmem.genMode.numtevstages + 1;
 	int numTexgen = bpmem.genMode.numtexgens;
 
+	bool per_pixel_depth = bpmem.ztex2.op != ZTEXTURE_DISABLE && !bpmem.zcontrol.early_ztest && bpmem.zmode.testenable;
+
 	char *p = text;
 	WRITE(p, "//Pixel Shader for TEV stages\n");
 	WRITE(p, "//%i TEV stages, %i texgens, XXX IND stages\n",
@@ -507,138 +528,216 @@ const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType
 				nIndirectStagesUsed |= 1 << bpmem.tevind[i].bt;
 		}
 	}
-	// Declare samplers
 
-	if(ApiType != API_D3D11)
+	if (ApiType == API_OPENGL)
 	{
-		WRITE(p, "uniform sampler2D ");
+
+		// A function here
+		// Fmod implementation gleaned from Nvidia
+		// At http://http.developer.nvidia.com/Cg/fmod.html
+		WRITE(p, "float fmod( float x, float y )\n");
+		WRITE(p, "{\n");
+		WRITE(p, "\tfloat z = fract( abs( x / y) ) * abs( y );\n");
+		WRITE(p, "\treturn (x < 0) ? -z : z;\n");
+		WRITE(p, "}\n");
+
+		for (int i = 0; i < 8; ++i)
+			WRITE(p, "uniform sampler2D samp%d;\n", i);
 	}
 	else
 	{
-		WRITE(p, "sampler ");
-	}
+		// Declare samplers
+		if (ApiType != API_D3D11)
+		{
+			WRITE(p, "uniform sampler2D ");
+		}
+		else
+		{
+			WRITE(p, "sampler ");
+		}
 
-	bool bfirst = true;
-	for (int i = 0; i < 8; ++i)
-	{
-		WRITE(p, "%s samp%d : register(s%d)", bfirst?"":",", i, i);
-		bfirst = false;
-	}
-	WRITE(p, ";\n");
-	if(ApiType == API_D3D11)
-	{
-		WRITE(p, "Texture2D ");
-		bfirst = true;
+		bool bfirst = true;
 		for (int i = 0; i < 8; ++i)
 		{
-			WRITE(p, "%s Tex%d : register(t%d)", bfirst?"":",", i, i);
+			WRITE(p, "%s samp%d %s", bfirst?"":",", i, WriteRegister(ApiType, "s", i));
 			bfirst = false;
 		}
 		WRITE(p, ";\n");
+		if (ApiType == API_D3D11)
+		{
+			WRITE(p, "Texture2D ");
+			bfirst = true;
+			for (int i = 0; i < 8; ++i)
+			{
+				WRITE(p, "%s Tex%d : register(t%d)", bfirst?"":",", i, i);
+				bfirst = false;
+			}
+			WRITE(p, ";\n");
+		}
 	}
 
 	WRITE(p, "\n");
-
-	WRITE(p, "uniform float4 " I_COLORS"[4] : register(c%d);\n", C_COLORS);
-	WRITE(p, "uniform float4 " I_KCOLORS"[4] : register(c%d);\n", C_KCOLORS);
-	WRITE(p, "uniform float4 " I_ALPHA"[1] : register(c%d);\n", C_ALPHA);
-	WRITE(p, "uniform float4 " I_TEXDIMS"[8] : register(c%d);\n", C_TEXDIMS);
-	WRITE(p, "uniform float4 " I_ZBIAS"[2] : register(c%d);\n", C_ZBIAS);
-	WRITE(p, "uniform float4 " I_INDTEXSCALE"[2] : register(c%d);\n", C_INDTEXSCALE);
-	WRITE(p, "uniform float4 " I_INDTEXMTX"[6] : register(c%d);\n", C_INDTEXMTX);
-	WRITE(p, "uniform float4 " I_FOG"[3] : register(c%d);\n", C_FOG);
-
-	if(g_ActiveConfig.bEnablePixelLighting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
-	{
-		WRITE(p,"typedef struct { float4 col; float4 cosatt; float4 distatt; float4 pos; float4 dir; } Light;\n");
-		WRITE(p,"typedef struct { Light lights[8]; } s_" I_PLIGHTS";\n");
-		WRITE(p, "uniform s_" I_PLIGHTS" " I_PLIGHTS" : register(c%d);\n", C_PLIGHTS);
-		WRITE(p, "typedef struct { float4 C0, C1, C2, C3; } s_" I_PMATERIALS";\n");
-		WRITE(p, "uniform s_" I_PMATERIALS" " I_PMATERIALS" : register(c%d);\n", C_PMATERIALS);
-	}
-
-	WRITE(p, "void main(\n");
-	if(ApiType != API_D3D11)
+	if (g_ActiveConfig.backend_info.bSupportsGLSLUBO)
+		WRITE(p, "layout(std140) uniform PSBlock {\n");
+	
+	WRITE(p, "\t%sfloat4 " I_COLORS"[4] %s;\n", WriteLocation(ApiType), WriteRegister(ApiType, "c", C_COLORS));
+	WRITE(p, "\t%sfloat4 " I_KCOLORS"[4] %s;\n", WriteLocation(ApiType), WriteRegister(ApiType, "c", C_KCOLORS));
+	WRITE(p, "\t%sfloat4 " I_ALPHA"[1] %s;\n", WriteLocation(ApiType), WriteRegister(ApiType, "c", C_ALPHA));
+	WRITE(p, "\t%sfloat4 " I_TEXDIMS"[8] %s;\n", WriteLocation(ApiType), WriteRegister(ApiType, "c", C_TEXDIMS));
+	WRITE(p, "\t%sfloat4 " I_ZBIAS"[2] %s;\n", WriteLocation(ApiType), WriteRegister(ApiType, "c", C_ZBIAS));
+	WRITE(p, "\t%sfloat4 " I_INDTEXSCALE"[2] %s;\n", WriteLocation(ApiType),  WriteRegister(ApiType, "c", C_INDTEXSCALE));
+	WRITE(p, "\t%sfloat4 " I_INDTEXMTX"[6] %s;\n", WriteLocation(ApiType), WriteRegister(ApiType, "c", C_INDTEXMTX));
+	WRITE(p, "\t%sfloat4 " I_FOG"[3] %s;\n", WriteLocation(ApiType), WriteRegister(ApiType, "c", C_FOG));
+	
+	// For pixel lighting
+	WRITE(p, "\t%sfloat4 " I_PLIGHTS"[40] %s;\n", WriteLocation(ApiType), WriteRegister(ApiType, "c", C_PLIGHTS));
+	WRITE(p, "\t%sfloat4 " I_PMATERIALS"[4] %s;\n", WriteLocation(ApiType), WriteRegister(ApiType, "c", C_PMATERIALS));
+		
+	if (g_ActiveConfig.backend_info.bSupportsGLSLUBO)
+		WRITE(p, "};\n");
+
+	if (ApiType == API_OPENGL)
 	{
-		WRITE(p, "  out float4 ocol0 : COLOR0,%s%s\n  in float4 rawpos : %s,\n",
-			dstAlphaMode == DSTALPHA_DUAL_SOURCE_BLEND ? "\n  out float4 ocol1 : COLOR1," : "",
-			"\n  out float depth : DEPTH,",
-			ApiType & API_OPENGL ? "WPOS" : ApiType & API_D3D9_SM20 ? "POSITION" : "VPOS");
+		WRITE(p, "out float4 ocol0;\n");
+		if (dstAlphaMode == DSTALPHA_DUAL_SOURCE_BLEND)
+			WRITE(p, "out float4 ocol1;\n");
+		
+		if (per_pixel_depth)
+			WRITE(p, "#define depth gl_FragDepth\n");
+		WRITE(p, "float4 rawpos = gl_FragCoord;\n");
+
+		WRITE(p, "VARYIN float4 colors_02;\n");
+		WRITE(p, "VARYIN float4 colors_12;\n");
+		WRITE(p, "float4 colors_0 = colors_02;\n");
+		WRITE(p, "float4 colors_1 = colors_12;\n");
+
+		// compute window position if needed because binding semantic WPOS is not widely supported
+				// Let's set up attributes
+		if (xfregs.numTexGen.numTexGens < 7)
+		{
+			for (int i = 0; i < 8; ++i)
+			{
+				WRITE(p, "VARYIN float3 uv%d_2;\n", i);
+				WRITE(p, "float3 uv%d = uv%d_2;\n", i, i);
+			}
+			WRITE(p, "VARYIN float4 clipPos_2;\n");
+			WRITE(p, "float4 clipPos = clipPos_2;\n");
+			if (g_ActiveConfig.bEnablePixelLighting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
+			{
+				WRITE(p, "VARYIN float4 Normal_2;\n");
+				WRITE(p, "float4 Normal = Normal_2;\n");
+			}
+		}
+		else
+		{
+			// wpos is in w of first 4 texcoords
+			if (g_ActiveConfig.bEnablePixelLighting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
+			{
+				for (int i = 0; i < 8; ++i)
+				{
+					WRITE(p, "VARYIN float4 uv%d_2;\n", i);
+					WRITE(p, "float4 uv%d = uv%d_2;\n", i, i);
+				}
+			}
+			else
+			{
+				for (unsigned int i = 0; i < xfregs.numTexGen.numTexGens; ++i)
+				{
+					WRITE(p, "VARYIN float%d uv%d_2;\n", i < 4 ? 4 : 3 , i);
+					WRITE(p, "float%d uv%d = uv%d_2;\n", i < 4 ? 4 : 3 , i, i);
+				}
+			}
+			WRITE(p, "float4 clipPos;\n");
+		}
+		WRITE(p, "void main()\n{\n");
 	}
 	else
 	{
-		WRITE(p, "  out float4 ocol0 : SV_Target0,%s%s\n  in float4 rawpos : SV_Position,\n",
-			dstAlphaMode == DSTALPHA_DUAL_SOURCE_BLEND ? "\n  out float4 ocol1 : SV_Target1," : "",
-			"\n  out float depth : SV_Depth,");
-	}
+		WRITE(p, "void main(\n");
+		if (ApiType != API_D3D11)
+		{
+			WRITE(p, "  out float4 ocol0 : COLOR0,%s%s\n  in float4 rawpos : %s,\n",
+				dstAlphaMode == DSTALPHA_DUAL_SOURCE_BLEND ? "\n  out float4 ocol1 : COLOR1," : "",
+				per_pixel_depth ? "\n  out float depth : DEPTH," : "",
+				ApiType & API_D3D9_SM20 ? "POSITION" : "VPOS");
+		}
+		else
+		{
+			WRITE(p, "  out float4 ocol0 : SV_Target0,%s%s\n  in float4 rawpos : SV_Position,\n",
+				dstAlphaMode == DSTALPHA_DUAL_SOURCE_BLEND ? "\n  out float4 ocol1 : SV_Target1," : "",
+				per_pixel_depth ? "\n  out float depth : SV_Depth," : "");
+		}
 
-	WRITE(p, "  in float4 colors_0 : COLOR0,\n");
-	WRITE(p, "  in float4 colors_1 : COLOR1");
+		WRITE(p, "  in float4 colors_0 : COLOR0,\n");
+		WRITE(p, "  in float4 colors_1 : COLOR1");
 
-	// compute window position if needed because binding semantic WPOS is not widely supported
-	if (numTexgen < 7)
-	{
-		for (int i = 0; i < numTexgen; ++i)
-			WRITE(p, ",\n  in float3 uv%d : TEXCOORD%d", i, i);
-		WRITE(p, ",\n  in float4 clipPos : TEXCOORD%d", numTexgen);
-		if(g_ActiveConfig.bEnablePixelLighting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
-			WRITE(p, ",\n  in float4 Normal : TEXCOORD%d", numTexgen + 1);
-	}
-	else
-	{
-		// wpos is in w of first 4 texcoords
-		if(g_ActiveConfig.bEnablePixelLighting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
+		// compute window position if needed because binding semantic WPOS is not widely supported
+		if (numTexgen < 7)
 		{
-			for (int i = 0; i < 8; ++i)
-				WRITE(p, ",\n  in float4 uv%d : TEXCOORD%d", i, i);
+			for (int i = 0; i < numTexgen; ++i)
+				WRITE(p, ",\n  in float3 uv%d : TEXCOORD%d", i, i);
+			WRITE(p, ",\n  in float4 clipPos : TEXCOORD%d", numTexgen);
+			if(g_ActiveConfig.bEnablePixelLighting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
+				WRITE(p, ",\n  in float4 Normal : TEXCOORD%d", numTexgen + 1);
+			WRITE(p, "        ) {\n");
 		}
 		else
 		{
-			for (unsigned int i = 0; i < xfregs.numTexGen.numTexGens; ++i)
-				WRITE(p, ",\n  in float%d uv%d : TEXCOORD%d", i < 4 ? 4 : 3 , i, i);
+			// wpos is in w of first 4 texcoords
+			if(g_ActiveConfig.bEnablePixelLighting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
+			{
+				for (int i = 0; i < 8; ++i)
+					WRITE(p, ",\n  in float4 uv%d : TEXCOORD%d", i, i);
+			}
+			else
+			{
+				for (unsigned int i = 0; i < xfregs.numTexGen.numTexGens; ++i)
+					WRITE(p, ",\n  in float%d uv%d : TEXCOORD%d", i < 4 ? 4 : 3 , i, i);
+			}
+			WRITE(p, "        ) {\n");
+			WRITE(p, "\tfloat4 clipPos = float4(0.0f, 0.0f, 0.0f, 0.0f);");
 		}
 	}
-	WRITE(p, "        ) {\n");
 
 	WRITE(p, "  float4 c0 = " I_COLORS"[1], c1 = " I_COLORS"[2], c2 = " I_COLORS"[3], prev = float4(0.0f, 0.0f, 0.0f, 0.0f), textemp = float4(0.0f, 0.0f, 0.0f, 0.0f), rastemp = float4(0.0f, 0.0f, 0.0f, 0.0f), konsttemp = float4(0.0f, 0.0f, 0.0f, 0.0f);\n"
 			"  float3 comp16 = float3(1.0f, 255.0f, 0.0f), comp24 = float3(1.0f, 255.0f, 255.0f*255.0f);\n"
-			"  float4 alphabump=float4(0.0f,0.0f,0.0f,0.0f);\n"
+			"  float alphabump=0.0f;\n"
 			"  float3 tevcoord=float3(0.0f, 0.0f, 0.0f);\n"
 			"  float2 wrappedcoord=float2(0.0f,0.0f), tempcoord=float2(0.0f,0.0f);\n"
 			"  float4 cc0=float4(0.0f,0.0f,0.0f,0.0f), cc1=float4(0.0f,0.0f,0.0f,0.0f);\n"
 			"  float4 cc2=float4(0.0f,0.0f,0.0f,0.0f), cprev=float4(0.0f,0.0f,0.0f,0.0f);\n"
 			"  float4 crastemp=float4(0.0f,0.0f,0.0f,0.0f),ckonsttemp=float4(0.0f,0.0f,0.0f,0.0f);\n\n");
 
-	if(g_ActiveConfig.bEnablePixelLighting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
+	if (g_ActiveConfig.bEnablePixelLighting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
 	{
 		if (xfregs.numTexGen.numTexGens < 7)
 		{
-			WRITE(p,"float3 _norm0 = normalize(Normal.xyz);\n\n");
-			WRITE(p,"float3 pos = float3(clipPos.x,clipPos.y,Normal.w);\n");
+			WRITE(p,"\tfloat3 _norm0 = normalize(Normal.xyz);\n\n");
+			WRITE(p,"\tfloat3 pos = float3(clipPos.x,clipPos.y,Normal.w);\n");
 		}
 		else
 		{
-			WRITE(p,"  float3 _norm0 = normalize(float3(uv4.w,uv5.w,uv6.w));\n\n");
-			WRITE(p,"float3 pos = float3(uv0.w,uv1.w,uv7.w);\n");
+			WRITE(p,"\tfloat3 _norm0 = normalize(float3(uv4.w,uv5.w,uv6.w));\n\n");
+			WRITE(p,"\tfloat3 pos = float3(uv0.w,uv1.w,uv7.w);\n");
 		}
 
 
-		WRITE(p, "float4 mat, lacc;\n"
-		"float3 ldir, h;\n"
-		"float dist, dist2, attn;\n");
+		WRITE(p, "\tfloat4 mat, lacc;\n"
+		"\tfloat3 ldir, h;\n"
+		"\tfloat dist, dist2, attn;\n");
 
 		p = GenerateLightingShader(p, components, I_PMATERIALS, I_PLIGHTS, "colors_", "colors_");
 	}
 
 	if (numTexgen < 7)
-		WRITE(p, "clipPos = float4(rawpos.x, rawpos.y, clipPos.z, clipPos.w);\n");
+		WRITE(p, "\tclipPos = float4(rawpos.x, rawpos.y, clipPos.z, clipPos.w);\n");
 	else
-		WRITE(p, "float4 clipPos = float4(rawpos.x, rawpos.y, uv2.w, uv3.w);\n");
+		WRITE(p, "\tclipPos = float4(rawpos.x, rawpos.y, uv2.w, uv3.w);\n");
 
 	// HACK to handle cases where the tex gen is not enabled
 	if (numTexgen == 0)
 	{
-		WRITE(p, "float3 uv0 = float3(0.0f, 0.0f, 0.0f);\n");
+		WRITE(p, "\tfloat3 uv0 = float3(0.0f, 0.0f, 0.0f);\n");
 	}
 	else
 	{
@@ -647,8 +746,8 @@ const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType
 			// optional perspective divides
 			if (xfregs.texMtxInfo[i].projection == XF_TEXPROJ_STQ)
 			{
-				WRITE(p, "if (uv%d.z)", i);
-				WRITE(p, "	uv%d.xy = uv%d.xy / uv%d.z;\n", i, i, i);
+				WRITE(p, "\tif (uv%d.z != 0.0f)", i);
+				WRITE(p, "\t\tuv%d.xy = uv%d.xy / uv%d.z;\n", i, i, i);
 			}
 
 			WRITE(p, "uv%d.xy = uv%d.xy * " I_TEXDIMS"[%d].zw;\n", i, i, i);
@@ -656,16 +755,16 @@ const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType
 	}
 
 	// indirect texture map lookup
-	for(u32 i = 0; i < bpmem.genMode.numindstages; ++i)
+	for (u32 i = 0; i < bpmem.genMode.numindstages; ++i)
 	{
 		if (nIndirectStagesUsed & (1<<i))
 		{
 			int texcoord = bpmem.tevindref.getTexCoord(i);
 
 			if (texcoord < numTexgen)
-				WRITE(p, "tempcoord = uv%d.xy * " I_INDTEXSCALE"[%d].%s;\n", texcoord, i/2, (i&1)?"zw":"xy");
+				WRITE(p, "\ttempcoord = uv%d.xy * " I_INDTEXSCALE"[%d].%s;\n", texcoord, i/2, (i&1)?"zw":"xy");
 			else
-				WRITE(p, "tempcoord = float2(0.0f, 0.0f);\n");
+				WRITE(p, "\ttempcoord = float2(0.0f, 0.0f);\n");
 
 			char buffer[32];
 			sprintf(buffer, "float3 indtex%d", i);
@@ -686,36 +785,42 @@ const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType
 	for (int i = 0; i < numStages; i++)
 		WriteStage(p, i, ApiType); //build the equation for this stage
 
-	if(numStages)
+	if (numStages)
 	{
 		// The results of the last texenv stage are put onto the screen,
 		// regardless of the used destination register
 		if(bpmem.combiners[numStages - 1].colorC.dest != 0)
 		{
 			bool retrieveFromAuxRegister = !RegisterStates[bpmem.combiners[numStages - 1].colorC.dest].ColorNeedOverflowControl && RegisterStates[bpmem.combiners[numStages - 1].colorC.dest].AuxStored;
-			WRITE(p, "prev.rgb = %s%s;\n", retrieveFromAuxRegister ? "c" : "" , tevCOutputTable[bpmem.combiners[numStages - 1].colorC.dest]);
+			WRITE(p, "\tprev.rgb = %s%s;\n", retrieveFromAuxRegister ? "c" : "" , tevCOutputTable[bpmem.combiners[numStages - 1].colorC.dest]);
 			RegisterStates[0].ColorNeedOverflowControl = RegisterStates[bpmem.combiners[numStages - 1].colorC.dest].ColorNeedOverflowControl;
 		}
 		if(bpmem.combiners[numStages - 1].alphaC.dest != 0)
 		{
 			bool retrieveFromAuxRegister = !RegisterStates[bpmem.combiners[numStages - 1].alphaC.dest].AlphaNeedOverflowControl && RegisterStates[bpmem.combiners[numStages - 1].alphaC.dest].AuxStored;
-			WRITE(p, "prev.a = %s%s;\n", retrieveFromAuxRegister ? "c" : "" , tevAOutputTable[bpmem.combiners[numStages - 1].alphaC.dest]);
+			WRITE(p, "\tprev.a = %s%s;\n", retrieveFromAuxRegister ? "c" : "" , tevAOutputTable[bpmem.combiners[numStages - 1].alphaC.dest]);
 			RegisterStates[0].AlphaNeedOverflowControl = RegisterStates[bpmem.combiners[numStages - 1].alphaC.dest].AlphaNeedOverflowControl;
 		}
 	}
 	// emulation of unsigned 8 overflow when casting if needed
 	if(RegisterStates[0].AlphaNeedOverflowControl || RegisterStates[0].ColorNeedOverflowControl)
-		WRITE(p, "prev = frac(prev * (255.0f/256.0f)) * (256.0f/255.0f);\n");
+		WRITE(p, "\tprev = frac(prev * (255.0f/256.0f)) * (256.0f/255.0f);\n");
 
 	AlphaTest::TEST_RESULT Pretest = bpmem.alpha_test.TestResult();
 	if (Pretest == AlphaTest::UNDETERMINED)
-		WriteAlphaTest(p, ApiType, dstAlphaMode);
+		WriteAlphaTest(p, ApiType, dstAlphaMode, per_pixel_depth);
 
+	
 	// the screen space depth value = far z + (clip z / clip w) * z range
-	WRITE(p, "float zCoord = " I_ZBIAS"[1].x + (clipPos.z / clipPos.w) * " I_ZBIAS"[1].y;\n");
+	if(ApiType == API_OPENGL || ApiType == API_D3D11)
+		WRITE(p, "float zCoord = rawpos.z;\n");
+	else
+		// dx9 doesn't support 4 component position, so we have to calculate it again
+		WRITE(p, "float zCoord = " I_ZBIAS"[1].x + (clipPos.z / clipPos.w) * " I_ZBIAS"[1].y;\n");
 
-	// Note: depth textures are disabled if early depth test is enabled
-	if (bpmem.ztex2.op != ZTEXTURE_DISABLE && !bpmem.zcontrol.early_ztest && bpmem.zmode.testenable)
+	// depth texture can safely be ignored if the result won't be written to the depth buffer (early_ztest) and isn't used for fog either
+	bool skip_ztexture = !per_pixel_depth && !bpmem.fog.c_proj_fsel.fsel;
+	if (bpmem.ztex2.op != ZTEXTURE_DISABLE && !skip_ztexture)
 	{
 		// use the texture input of the last texture stage (textemp), hopefully this has been read and is in correct format...
 		WRITE(p, "zCoord = dot(" I_ZBIAS"[0].xyzw, textemp.xyzw) + " I_ZBIAS"[1].w %s;\n",
@@ -725,15 +830,18 @@ const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType
 		WRITE(p, "zCoord = zCoord * (16777215.0f/16777216.0f);\n");
 		WRITE(p, "zCoord = frac(zCoord);\n");
 		WRITE(p, "zCoord = zCoord * (16777216.0f/16777215.0f);\n");
+
+		// Note: depth texture out put is only written to depth buffer if late depth test is used
+		if (per_pixel_depth)
+			WRITE(p, "depth = zCoord;\n");
 	}
-	WRITE(p, "depth = zCoord;\n");
 
 	if (dstAlphaMode == DSTALPHA_ALPHA_PASS)
-		WRITE(p, "  ocol0 = float4(prev.rgb, " I_ALPHA"[0].a);\n");
+		WRITE(p, "\tocol0 = float4(prev.rgb, " I_ALPHA"[0].a);\n");
 	else
 	{
 		WriteFog(p);
-		WRITE(p, "  ocol0 = prev;\n");
+		WRITE(p, "\tocol0 = prev;\n");
 	}
 
 	// On D3D11, use dual-source color blending to perform dst alpha in a
@@ -741,9 +849,9 @@ const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType
 	if (dstAlphaMode == DSTALPHA_DUAL_SOURCE_BLEND)
 	{
 		// Colors will be blended against the alpha from ocol1...
-		WRITE(p, "  ocol1 = ocol0;\n");
+		WRITE(p, "\tocol1 = prev;\n");
 		// ...and the alpha from ocol0 will be written to the framebuffer.
-		WRITE(p, "  ocol0.a = " I_ALPHA"[0].a;\n");
+		WRITE(p, "\tocol0.a = " I_ALPHA"[0].a;\n");	
 	}
 	
 	WRITE(p, "}\n");
@@ -852,10 +960,10 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 				WRITE(p, "float2 indtevtrans%d = " I_INDTEXMTX"[%d].ww * uv%d.xy * indtevcrd%d.yy;\n", n, mtxidx, texcoord, n);
 			}
 			else
-				WRITE(p, "float2 indtevtrans%d = 0;\n", n);
+				WRITE(p, "float2 indtevtrans%d = float2(0.0f, 0.0f);\n", n);
 		}
 		else
-			WRITE(p, "float2 indtevtrans%d = 0;\n", n);
+			WRITE(p, "float2 indtevtrans%d = float2(0.0f, 0.0f);\n", n);
 
 		// ---------
 		// Wrapping
@@ -901,10 +1009,10 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 
 	if (bpmem.tevorders[n/2].getEnable(n&1))
 	{
-		if(!bHasIndStage)
+		if (!bHasIndStage)
 		{
 			// calc tevcord
-			if(bHasTexCoord)
+			if (bHasTexCoord)
 				WRITE(p, "tevcoord.xy = uv%d.xy;\n", texcoord);
 			else
 				WRITE(p, "tevcoord.xy = float2(0.0f, 0.0f);\n");
@@ -924,7 +1032,7 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 		int kc = bpmem.tevksel[n / 2].getKC(n & 1);
 		int ka = bpmem.tevksel[n / 2].getKA(n & 1);
 		WRITE(p, "konsttemp = float4(%s, %s);\n", tevKSelTableC[kc], tevKSelTableA[ka]);
-		if(kc > 7 || ka > 7)
+		if (kc > 7 || ka > 7)
 		{
 			WRITE(p, "ckonsttemp = frac(konsttemp * (255.0f/256.0f)) * (256.0f/255.0f);\n");
 		}
@@ -1023,7 +1131,7 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 		if (cc.shift > TEVSCALE_1)
 			WRITE(p, "%s*(", tevScaleTable[cc.shift]);
 
-		if(!(cc.d == TEVCOLORARG_ZERO && cc.op == TEVOP_ADD))
+		if (!(cc.d == TEVCOLORARG_ZERO && cc.op == TEVOP_ADD))
 			WRITE(p, "%s%s", tevCInputTable[cc.d], tevOpTable[cc.op]);
 
 		if (cc.a == cc.b)
@@ -1073,7 +1181,7 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 		if (ac.shift > TEVSCALE_1)
 			WRITE(p, "%s*(", tevScaleTable[ac.shift]);
 
-		if(!(ac.d == TEVALPHAARG_ZERO && ac.op == TEVOP_ADD))
+		if (!(ac.d == TEVALPHAARG_ZERO && ac.op == TEVOP_ADD))
 			WRITE(p, "%s.a%s", tevAInputTable[ac.d], tevOpTable[ac.op]);
 
 		if (ac.a == ac.b)
@@ -1089,7 +1197,7 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 
 		WRITE(p, "%s",tevBiasTable[ac.bias]);
 
-		if (ac.shift>0)
+		if (ac.shift > 0)
 			WRITE(p, ")");
 
 	}
@@ -1114,7 +1222,7 @@ void SampleTexture(char *&p, const char *destination, const char *texcoords, con
 	if (ApiType == API_D3D11)
 		WRITE(p, "%s=Tex%d.Sample(samp%d,%s.xy * " I_TEXDIMS"[%d].xy).%s;\n", destination, texmap,texmap, texcoords, texmap, texswap);
 	else
-		WRITE(p, "%s=tex2D(samp%d,%s.xy * " I_TEXDIMS"[%d].xy).%s;\n", destination, texmap, texcoords, texmap, texswap);
+		WRITE(p, "%s=%s(samp%d,%s.xy * " I_TEXDIMS"[%d].xy).%s;\n", destination, ApiType == API_OPENGL ? "texture" : "tex2D", texmap, texcoords, texmap, texswap);
 }
 
 static const char *tevAlphaFuncsTable[] =
@@ -1137,7 +1245,7 @@ static const char *tevAlphaFunclogicTable[] =
 	" == "  // xnor
 };
 
-static void WriteAlphaTest(char *&p, API_TYPE ApiType,DSTALPHA_MODE dstAlphaMode)
+static void WriteAlphaTest(char *&p, API_TYPE ApiType,DSTALPHA_MODE dstAlphaMode, bool per_pixel_depth)
 {
 	static const char *alphaRef[2] =
 	{
@@ -1145,8 +1253,9 @@ static void WriteAlphaTest(char *&p, API_TYPE ApiType,DSTALPHA_MODE dstAlphaMode
 		I_ALPHA"[0].g"
 	};	
 
+
 	// using discard then return works the same in cg and dx9 but not in dx11
-	WRITE(p, "if(!( ");
+	WRITE(p, "\tif(!( ");
 
 	int compindex = bpmem.alpha_test.comp0;
 	WRITE(p, tevAlphaFuncsTable[compindex],alphaRef[0]);//lookup the first component from the alpha function table
@@ -1157,10 +1266,11 @@ static void WriteAlphaTest(char *&p, API_TYPE ApiType,DSTALPHA_MODE dstAlphaMode
 	WRITE(p, tevAlphaFuncsTable[compindex],alphaRef[1]);//lookup the second component from the alpha function table
 	WRITE(p, ")) {\n");
 
-	WRITE(p, "ocol0 = 0;\n");
+	WRITE(p, "\t\tocol0 = float4(0.0f, 0.0f, 0.0f, 0.0f);\n");
 	if (dstAlphaMode == DSTALPHA_DUAL_SOURCE_BLEND)
-		WRITE(p, "ocol1 = 0;\n");
-	WRITE(p, "depth = 1.f;\n");
+		WRITE(p, "\t\tocol1 = float4(0.0f, 0.0f, 0.0f, 0.0f);\n");
+	if(per_pixel_depth)
+		WRITE(p, "depth = 1.f;\n");
 
 	// HAXX: zcomploc (aka early_ztest) is a way to control whether depth test is done before
 	// or after texturing and alpha test. PC GPUs have no way to support this
@@ -1174,9 +1284,9 @@ static void WriteAlphaTest(char *&p, API_TYPE ApiType,DSTALPHA_MODE dstAlphaMode
 	// we don't have a choice.
 	if (!(bpmem.zcontrol.early_ztest && bpmem.zmode.updateenable))
 	{
-		WRITE(p, "discard;\n");
+		WRITE(p, "\t\tdiscard;\n");
 		if (ApiType != API_D3D11)
-			WRITE(p, "return;\n");
+			WRITE(p, "\t\treturn;\n");
 	}
 
 	WRITE(p, "}\n");
@@ -1188,52 +1298,51 @@ static const char *tevFogFuncsTable[] =
 	"",																//?
 	"",																//Linear
 	"",																//?
-	"  fog = 1.0f - pow(2.0f, -8.0f * fog);\n",						//exp
-	"  fog = 1.0f - pow(2.0f, -8.0f * fog * fog);\n",					//exp2
-	"  fog = pow(2.0f, -8.0f * (1.0f - fog));\n",						//backward exp
-	"  fog = 1.0f - fog;\n   fog = pow(2.0f, -8.0f * fog * fog);\n"	//backward exp2
+	"\tfog = 1.0f - pow(2.0f, -8.0f * fog);\n",						//exp
+	"\tfog = 1.0f - pow(2.0f, -8.0f * fog * fog);\n",					//exp2
+	"\tfog = pow(2.0f, -8.0f * (1.0f - fog));\n",						//backward exp
+	"\tfog = 1.0f - fog;\n   fog = pow(2.0f, -8.0f * fog * fog);\n"	//backward exp2
 };
 
 static void WriteFog(char *&p)
 {
-	if(bpmem.fog.c_proj_fsel.fsel == 0)return;//no Fog
+	if (bpmem.fog.c_proj_fsel.fsel == 0)
+		return; // no Fog
 
 	if (bpmem.fog.c_proj_fsel.proj == 0)
 	{
 		// perspective
 		// ze = A/(B - (Zs >> B_SHF)
-		WRITE (p, "  float ze = " I_FOG"[1].x / (" I_FOG"[1].y - (zCoord / " I_FOG"[1].w));\n");
+		WRITE (p, "\tfloat ze = " I_FOG"[1].x / (" I_FOG"[1].y - (zCoord / " I_FOG"[1].w));\n");
 	}
 	else
 	{
 		// orthographic
 		// ze = a*Zs	(here, no B_SHF)
-		WRITE (p, "  float ze = " I_FOG"[1].x * zCoord;\n");
+		WRITE (p, "\tfloat ze = " I_FOG"[1].x * zCoord;\n");
 	}
 
 	// x_adjust = sqrt((x-center)^2 + k^2)/k
 	// ze *= x_adjust
 	//this is complitly teorical as the real hard seems to use a table intead of calculate the values.
-	if(bpmem.fogRange.Base.Enabled)
+	if (bpmem.fogRange.Base.Enabled)
 	{
-		WRITE (p, "  float x_adjust = (2.0f * (clipPos.x / " I_FOG"[2].y)) - 1.0f - " I_FOG"[2].x;\n");
-		WRITE (p, "  x_adjust = sqrt(x_adjust * x_adjust + " I_FOG"[2].z * " I_FOG"[2].z) / " I_FOG"[2].z;\n");
-		WRITE (p, "  ze *= x_adjust;\n");
+		WRITE (p, "\tfloat x_adjust = (2.0f * (clipPos.x / " I_FOG"[2].y)) - 1.0f - " I_FOG"[2].x;\n");
+		WRITE (p, "\tx_adjust = sqrt(x_adjust * x_adjust + " I_FOG"[2].z * " I_FOG"[2].z) / " I_FOG"[2].z;\n");
+		WRITE (p, "\tze *= x_adjust;\n");
 	}
 
-	WRITE (p, "  float fog = saturate(ze - " I_FOG"[1].z);\n");
+	WRITE (p, "\tfloat fog = saturate(ze - " I_FOG"[1].z);\n");
 
-	if(bpmem.fog.c_proj_fsel.fsel > 3)
+	if (bpmem.fog.c_proj_fsel.fsel > 3)
 	{
 		WRITE(p, "%s", tevFogFuncsTable[bpmem.fog.c_proj_fsel.fsel]);
 	}
 	else
 	{
-		if(bpmem.fog.c_proj_fsel.fsel != 2)
+		if (bpmem.fog.c_proj_fsel.fsel != 2)
 			WARN_LOG(VIDEO, "Unknown Fog Type! %08x", bpmem.fog.c_proj_fsel.fsel);
 	}
 
-	WRITE(p, "  prev.rgb = lerp(prev.rgb," I_FOG"[0].rgb,fog);\n");
-
-
+	WRITE(p, "\tprev.rgb = lerp(prev.rgb, " I_FOG"[0].rgb, fog);\n");
 }