libctru 1.7.1

Used to configure whether the Z component of the normal map is used or if
the Z component is reconstructed based on the XY components of the normal map.

Makefile

@@ -9,8 +9,8 @@ endif
 include $(DEVKITARM)/3ds_rules
 
 export CITRO3D_MAJOR	:= 1
-export CITRO3D_MINOR	:= 6
+export CITRO3D_MINOR	:= 7
-export CITRO3D_PATCH	:= 0
+export CITRO3D_PATCH	:= 1
 
 VERSION	:=	$(CITRO3D_MAJOR).$(CITRO3D_MINOR).$(CITRO3D_PATCH)
 
@@ -31,11 +31,11 @@ INCLUDES	:=	include
 #---------------------------------------------------------------------------------
 ARCH	:=	-march=armv6k -mtune=mpcore -mfloat-abi=hard -mtp=soft
 
-CFLAGS	:=	-g -Wall -Werror -mword-relocations \
+CFLAGS	:=	-g -Wall -Wno-sizeof-array-div -Werror -mword-relocations \
 			-ffunction-sections -fdata-sections \
 			$(ARCH) $(BUILD_CFLAGS)
 
-CFLAGS	+=	$(INCLUDE) -DARM11 -D_3DS -DCITRO3D_BUILD
+CFLAGS	+=	$(INCLUDE) -D__3DS__ -DCITRO3D_BUILD
 
 CXXFLAGS	:= $(CFLAGS) -fno-rtti -fno-exceptions -std=gnu++11
 
@@ -114,7 +114,7 @@ debug:
 
 lib/libcitro3d.a : lib release $(SOURCES) $(INCLUDES)
 	@$(MAKE) BUILD=release OUTPUT=$(CURDIR)/$@ \
-	BUILD_CFLAGS="-DNDEBUG=1 -O2 -fomit-frame-pointer" \
+	BUILD_CFLAGS="-DNDEBUG=1 -O2 -fomit-frame-pointer -fno-math-errno" \
 	DEPSDIR=$(CURDIR)/release \
 	--no-print-directory -C release \
 	-f $(CURDIR)/Makefile

include/c3d/light.h

@@ -77,6 +77,14 @@ enum
 void C3D_LightEnvFresnel(C3D_LightEnv* env, GPU_FRESNELSEL selector);
 void C3D_LightEnvBumpMode(C3D_LightEnv* env, GPU_BUMPMODE mode);
 void C3D_LightEnvBumpSel(C3D_LightEnv* env, int texUnit);
+
+/**
+ * @brief Configures whether to use the z component of the normal map.
+ * @param[out] env    Pointer to light environment structure.
+ * @param[in]  enable false if the z component is reconstructed from the xy components
+ *                     of the normal map, true if the z component is taken from the normal map.
+ */
+void C3D_LightEnvBumpNormalZ(C3D_LightEnv *env, bool enable);
 void C3D_LightEnvShadowMode(C3D_LightEnv* env, u32 mode);
 void C3D_LightEnvShadowSel(C3D_LightEnv* env, int texUnit);
 void C3D_LightEnvClampHighlights(C3D_LightEnv* env, bool clamp);

include/c3d/maths.h

@@ -13,7 +13,12 @@
  * The one true circumference-to-radius ratio.
  * See http://tauday.com/tau-manifesto
  */
-#define M_TAU (2*M_PI)
+#define M_TAU (6.28318530717958647692528676655900576)
+
+// Define the legacy circle constant as well
+#ifndef M_PI
+#define M_PI (M_TAU/2)
+#endif
 
 /**
  * @brief Convert an angle from revolutions to radians

include/c3d/renderqueue.h

@@ -64,6 +64,11 @@ C3D_RenderTarget* C3D_RenderTargetCreateFromTex(C3D_Tex* tex, GPU_TEXFACE face,
 void C3D_RenderTargetDelete(C3D_RenderTarget* target);
 void C3D_RenderTargetSetOutput(C3D_RenderTarget* target, gfxScreen_t screen, gfx3dSide_t side, u32 transferFlags);
 
+static inline void C3D_RenderTargetDetachOutput(C3D_RenderTarget* target)
+{
+	C3D_RenderTargetSetOutput(NULL, target->screen, target->side, 0);
+}
+
 static inline void C3D_RenderTargetClear(C3D_RenderTarget* target, C3D_ClearBits clearBits, u32 clearColor, u32 clearDepth)
 {
 	C3D_FrameBufClear(&target->frameBuf, clearBits, clearColor, clearDepth);

include/c3d/texture.h

@@ -41,7 +41,7 @@ typedef struct
 	};
 } C3D_Tex;
 
-typedef struct ALIGN(8)
+typedef struct CTR_ALIGN(8)
 {
 	u16 width;
 	u16 height;

include/c3d/types.h

@@ -1,5 +1,5 @@
 #pragma once
-#ifdef _3DS
+#if defined(__3DS__) || defined(_3DS)
 #include <3ds.h>
 #else
 #include <stdbool.h>

include/tex3ds.h

@@ -38,7 +38,7 @@ extern "C" {
 #endif
 
 /** @brief Subtexture
- *  @note If top > bottom, the subtexture is rotated 1/4 revolution counter-clockwise
+ *  @note If top < bottom, the subtexture is rotated 1/4 revolution counter-clockwise
  */
 typedef struct Tex3DS_SubTexture
 {

source/base.c

@@ -42,10 +42,12 @@ static void C3Di_AptEventHook(APT_HookType hookType, C3D_UNUSED void* param)
 		case APTHOOK_ONSUSPEND:
 		{
 			C3Di_RenderQueueWaitDone();
+			C3Di_RenderQueueDisableVBlank();
 			break;
 		}
 		case APTHOOK_ONRESTORE:
 		{
+			C3Di_RenderQueueEnableVBlank();
 			ctx->flags |= C3DiF_AttrInfo | C3DiF_BufInfo | C3DiF_Effect | C3DiF_FrameBuf
 				| C3DiF_Viewport | C3DiF_Scissor | C3DiF_Program | C3DiF_VshCode | C3DiF_GshCode
 				| C3DiF_TexAll | C3DiF_TexEnvBuf | C3DiF_TexEnvAll | C3DiF_LightEnv | C3DiF_Gas;

source/internal.h

@@ -124,6 +124,18 @@ static inline bool C3Di_TexIs2D(C3D_Tex* tex)
 	return !typeIsCube(C3D_TexGetType(tex));
 }
 
+static inline bool addrIsVRAM(const void* addr)
+{
+	u32 vaddr = (u32)addr;
+	return vaddr >= OS_VRAM_VADDR && vaddr < OS_VRAM_VADDR + OS_VRAM_SIZE;
+}
+
+static inline vramAllocPos addrGetVRAMBank(const void* addr)
+{
+	u32 vaddr = (u32)addr;
+	return vaddr < OS_VRAM_VADDR + OS_VRAM_SIZE/2 ? VRAM_ALLOC_A : VRAM_ALLOC_B;
+}
+
 void C3Di_UpdateContext(void);
 void C3Di_AttrInfoBind(C3D_AttrInfo* info);
 void C3Di_BufInfoBind(C3D_BufInfo* info);
@@ -144,3 +156,5 @@ bool C3Di_SplitFrame(u32** pBuf, u32* pSize);
 void C3Di_RenderQueueInit(void);
 void C3Di_RenderQueueExit(void);
 void C3Di_RenderQueueWaitDone(void);
+void C3Di_RenderQueueEnableVBlank(void);
+void C3Di_RenderQueueDisableVBlank(void);

source/lightenv.c

@@ -250,6 +250,14 @@ void C3D_LightEnvBumpSel(C3D_LightEnv* env, int texUnit)
 	env->flags |= C3DF_LightEnv_Dirty;
 }
 
+void C3D_LightEnvBumpNormalZ(C3D_LightEnv *env, bool usez) {
+	if (usez)
+		env->conf.config[0] |= BIT(30);
+	else
+		env->conf.config[0] &= ~BIT(30);
+	env->flags |= C3DF_LightEnv_Dirty;
+}
+
 void C3D_LightEnvShadowMode(C3D_LightEnv* env, u32 mode)
 {
 	mode &= 0xF<<16;

source/renderqueue.c

@@ -9,7 +9,7 @@ static C3D_RenderTarget *linkedTarget[3];
 static TickCounter gpuTime, cpuTime;
 
 static bool inFrame, inSafeTransfer, measureGpuTime;
-static bool needSwapTop, needSwapBot;
+static bool needSwapTop, needSwapBot, isTopStereo;
 static float framerate = 60.0f;
 static float framerateCounter[2] = { 60.0f, 60.0f };
 static u32 frameCounter[2];
@@ -61,12 +61,12 @@ static void onQueueFinish(gxCmdQueue_s* queue)
 	{
 		if (needSwapTop)
 		{
-			gfxConfigScreen(GFX_TOP, false);
+			gfxScreenSwapBuffers(GFX_TOP, isTopStereo);
 			needSwapTop = false;
 		}
 		if (needSwapBot)
 		{
-			gfxConfigScreen(GFX_BOTTOM, false);
+			gfxScreenSwapBuffers(GFX_BOTTOM, false);
 			needSwapBot = false;
 		}
 	}
@@ -99,12 +99,23 @@ static bool C3Di_WaitAndClearQueue(s64 timeout)
 	return true;
 }
 
+void C3Di_RenderQueueEnableVBlank(void)
+{
+	gspSetEventCallback(GSPGPU_EVENT_VBlank0, onVBlank0, NULL, false);
+	gspSetEventCallback(GSPGPU_EVENT_VBlank1, onVBlank1, NULL, false);
+}
+
+void C3Di_RenderQueueDisableVBlank(void)
+{
+	gspSetEventCallback(GSPGPU_EVENT_VBlank0, NULL, NULL, false);
+	gspSetEventCallback(GSPGPU_EVENT_VBlank1, NULL, NULL, false);
+}
+
 void C3Di_RenderQueueInit(void)
 {
 	C3D_Context* ctx = C3Di_GetContext();
 
-	gspSetEventCallback(GSPGPU_EVENT_VBlank0, onVBlank0, NULL, false);
+	C3Di_RenderQueueEnableVBlank();
-	gspSetEventCallback(GSPGPU_EVENT_VBlank1, onVBlank1, NULL, false);
 
 	GX_BindQueue(&ctx->gxQueue);
 	gxCmdQueueSetCallback(&ctx->gxQueue, onQueueFinish, NULL);
@@ -120,8 +131,7 @@ void C3Di_RenderQueueExit(void)
 	gxCmdQueueSetCallback(&C3Di_GetContext()->gxQueue, NULL, NULL);
 	GX_BindQueue(NULL);
 
-	gspSetEventCallback(GSPGPU_EVENT_VBlank0, NULL, NULL, false);
+	C3Di_RenderQueueDisableVBlank();
-	gspSetEventCallback(GSPGPU_EVENT_VBlank1, NULL, NULL, false);
 
 	for (i = 0; i < 3; i ++)
 		linkedTarget[i] = NULL;
@@ -207,6 +217,7 @@ void C3D_FrameEnd(u8 flags)
 
 	int i;
 	C3D_RenderTarget* target;
+	isTopStereo = false;
 	for (i = 2; i >= 0; i --)
 	{
 		target = linkedTarget[i];
@@ -215,7 +226,11 @@ void C3D_FrameEnd(u8 flags)
 		target->used = false;
 		C3D_FrameBufTransfer(&target->frameBuf, target->screen, target->side, target->transferFlags);
 		if (target->screen == GFX_TOP)
+		{
 			needSwapTop = true;
+			if (target->side == GFX_RIGHT)
+				isTopStereo = true;
+		}
 		else if (target->screen == GFX_BOTTOM)
 			needSwapBot = true;
 	}
@@ -269,7 +284,10 @@ C3D_RenderTarget* C3D_RenderTargetCreate(int width, int height, GPU_COLORBUF col
 	if (C3D_DEPTHTYPE_OK(depthFmt))
 	{
 		depthFmtReal = C3D_DEPTHTYPE_VAL(depthFmt);
-		depthBuf = vramAlloc(C3D_CalcDepthBufSize(width,height,depthFmtReal));
+		size_t depthSize = C3D_CalcDepthBufSize(width,height,depthFmtReal);
+		vramAllocPos vramBank = addrGetVRAMBank(colorBuf);
+		depthBuf = vramAllocAt(depthSize, vramBank ^ VRAM_ALLOC_ANY); // Attempt opposite bank first...
+		if (!depthBuf) depthBuf = vramAllocAt(depthSize, vramBank); // ... if that fails, attempt same bank
 		if (!depthBuf) goto _fail1;
 	}
 
@@ -298,6 +316,7 @@ _fail0:
 
 C3D_RenderTarget* C3D_RenderTargetCreateFromTex(C3D_Tex* tex, GPU_TEXFACE face, int level, C3D_DEPTHTYPE depthFmt)
 {
+	if (!addrIsVRAM(tex->data)) return NULL; // Render targets must be in VRAM
 	C3D_RenderTarget* target = C3Di_RenderTargetNew();
 	if (!target) return NULL;
 
@@ -307,7 +326,10 @@ C3D_RenderTarget* C3D_RenderTargetCreateFromTex(C3D_Tex* tex, GPU_TEXFACE face,
 	if (C3D_DEPTHTYPE_OK(depthFmt))
 	{
 		GPU_DEPTHBUF depthFmtReal = C3D_DEPTHTYPE_VAL(depthFmt);
-		void* depthBuf = vramAlloc(C3D_CalcDepthBufSize(fb->width,fb->height,depthFmtReal));
+		size_t depthSize = C3D_CalcDepthBufSize(fb->width,fb->height,depthFmtReal);
+		vramAllocPos vramBank = addrGetVRAMBank(tex->data);
+		void* depthBuf = vramAllocAt(depthSize, vramBank ^ VRAM_ALLOC_ANY); // Attempt opposite bank first...
+		if (!depthBuf) depthBuf = vramAllocAt(depthSize, vramBank); // ... if that fails, attempt same bank
 		if (!depthBuf)
 		{
 			free(target);
@@ -340,6 +362,9 @@ void C3D_RenderTargetDelete(C3D_RenderTarget* target)
 {
 	if (inFrame)
 		svcBreak(USERBREAK_PANIC); // Shouldn't happen.
+	if (target->linked)
+		C3D_RenderTargetDetachOutput(target);
+	else
 		C3Di_WaitAndClearQueue(-1);
 	C3Di_RenderTargetDestroy(target);
 }
@@ -350,12 +375,19 @@ void C3D_RenderTargetSetOutput(C3D_RenderTarget* target, gfxScreen_t screen, gfx
 	if (screen==GFX_BOTTOM) id = 2;
 	else if (side==GFX_RIGHT) id = 1;
 	if (linkedTarget[id])
+	{
 		linkedTarget[id]->linked = false;
+		if (!inFrame)
+			C3Di_WaitAndClearQueue(-1);
+	}
 	linkedTarget[id] = target;
+	if (target)
+	{
 		target->linked = true;
 		target->transferFlags = transferFlags;
 		target->screen = screen;
 		target->side = side;
+	}
 }
 
 static void C3Di_SafeDisplayTransfer(u32* inadr, u32 indim, u32* outadr, u32 outdim, u32 flags)

source/texture.c

@@ -30,12 +30,6 @@ static inline size_t fmtSize(GPU_TEXCOLOR fmt)
 	}
 }
 
-static inline bool addrIsVRAM(const void* addr)
-{
-	u32 vaddr = (u32)addr;
-	return vaddr >= 0x1F000000 && vaddr < 0x1F600000;
-}
-
 static inline bool checkTexSize(u32 size)
 {
 	if (size < 8 || size > 1024)

test/3ds/Makefile

@@ -47,10 +47,10 @@ ICON            :=
 ARCH     := -march=armv6k -mtune=mpcore -mfloat-abi=hard -mtp=soft
 
 CFLAGS   := -g -Wall -O3 -mword-relocations \
-            -fomit-frame-pointer -ffunction-sections \
+            -ffunction-sections \
             $(ARCH)
 
-CFLAGS   +=  $(INCLUDE) -DARM11 -D_3DS
+CFLAGS   +=  $(INCLUDE) -D__3DS__
 
 CXXFLAGS := $(CFLAGS) -fno-rtti -std=gnu++11