From 9425edc406832fd98f0f1262a5ab1ea5de16ab39 Mon Sep 17 00:00:00 2001
From: yellows8 <yellows8@users.noreply.github.com>
Date: Fri, 29 Aug 2014 16:19:58 -0400
Subject: [PATCH 01/14] Added HID event code, based on the GSP event code.
 Added event id check in gspWaitForEvent().

---
 libctru/include/3ds/services/hid.h | 17 ++++++++++++++++-
 libctru/source/services/gsp.c      |  2 ++
 libctru/source/services/hid.c      | 23 +++++++++++++++++++++--
 3 files changed, 39 insertions(+), 3 deletions(-)

diff --git a/libctru/include/3ds/services/hid.h b/libctru/include/3ds/services/hid.h
index 17e13ab..b39898f 100644
--- a/libctru/include/3ds/services/hid.h
+++ b/libctru/include/3ds/services/hid.h
@@ -1,5 +1,7 @@
 #pragma once
 
+//See also: http://3dbrew.org/wiki/HID_Services http://3dbrew.org/wiki/HID_Shared_Memory
+
 #define HID_SHAREDMEM_DEFAULT (0x10000000)
 
 typedef enum
@@ -39,6 +41,17 @@ typedef struct
 	s16 dx, dy;
 } circlePosition;
 
+typedef enum
+{
+	HIDEVENT_PAD0 = 0, //"Event signaled by HID-module, when the sharedmem+0(PAD/circle-pad)/+0xA8(touch-screen) region was updated."
+	HIDEVENT_PAD1, //"Event signaled by HID-module, when the sharedmem+0(PAD/circle-pad)/+0xA8(touch-screen) region was updated."
+	HIDEVENT_Accel, //"Event signaled by HID-module, when the sharedmem accelerometer state was updated."
+	HIDEVENT_Gyro, //"Event signaled by HID-module, when the sharedmem gyroscope state was updated."
+	HIDEVENT_DebugPad, //"Event signaled by HID-module, when the sharedmem DebugPad state was updated."
+
+	HIDEVENT_MAX, // used to know how many events there are
+} HID_Event;
+
 extern Handle hidMemHandle;
 extern vu32* hidSharedMem;
 
@@ -52,6 +65,8 @@ u32 hidKeysUp();
 void hidTouchRead(touchPosition* pos);
 void hidCircleRead(circlePosition* pos);
 
+void hidWaitForEvent(HID_Event id, bool nextEvent);
+
 // libnds compatibility defines
 #define scanKeys   hidScanInput
 #define keysHeld   hidKeysHeld
@@ -60,7 +75,7 @@ void hidCircleRead(circlePosition* pos);
 #define touchRead  hidTouchRead
 #define circleRead hidCircleRead
 
-Result HIDUSER_GetSharedMem(Handle* outMemHandle);
+Result HIDUSER_GetHandles(Handle* outMemHandle, Handle *eventpad0, Handle *eventpad1, Handle *eventaccel, Handle *eventgyro, Handle *eventdebugpad);
 Result HIDUSER_EnableAccelerometer();
 Result HIDUSER_DisableAccelerometer();
 Result HIDUSER_EnableGyroscope();
diff --git a/libctru/source/services/gsp.c b/libctru/source/services/gsp.c
index bd76035..7c64019 100644
--- a/libctru/source/services/gsp.c
+++ b/libctru/source/services/gsp.c
@@ -68,6 +68,8 @@ void gspExitEventHandler()
 
 void gspWaitForEvent(GSP_Event id, bool nextEvent)
 {
+	if(id>=GSPEVENT_MAX)return;
+
 	if (nextEvent)
 		svcClearEvent(gspEvents[id]);
 	svcWaitSynchronization(gspEvents[id], U64_MAX);
diff --git a/libctru/source/services/hid.c b/libctru/source/services/hid.c
index 1265ec9..aaff8e1 100644
--- a/libctru/source/services/hid.c
+++ b/libctru/source/services/hid.c
@@ -8,6 +8,8 @@
 Handle hidHandle;
 Handle hidMemHandle;
 
+Handle hidEvents[5];
+
 vu32* hidSharedMem;
 
 static u32 kOld, kHeld, kDown, kUp;
@@ -24,7 +26,7 @@ Result hidInit(u32* sharedMem)
 	if((ret=srvGetServiceHandle(&hidHandle, "hid:USER")))return ret;
 
 	// Get sharedmem handle.
-	if((ret=HIDUSER_GetSharedMem(&hidMemHandle))) goto cleanup1;
+	if((ret=HIDUSER_GetHandles(&hidMemHandle, &hidEvents[HIDEVENT_PAD0], &hidEvents[HIDEVENT_PAD1], &hidEvents[HIDEVENT_Accel], &hidEvents[HIDEVENT_Gyro], &hidEvents[HIDEVENT_DebugPad]))) goto cleanup1;
 
 	// Map HID shared memory at addr "sharedMem".
 	hidSharedMem=sharedMem;
@@ -49,6 +51,17 @@ void hidExit()
 	svcCloseHandle(hidHandle);
 }
 
+void hidWaitForEvent(HID_Event id, bool nextEvent)
+{
+	if(id>=HIDEVENT_MAX)return;
+
+	if (nextEvent)
+		svcClearEvent(hidEvents[id]);
+	svcWaitSynchronization(hidEvents[id], U64_MAX);
+	if (!nextEvent)
+		svcClearEvent(hidEvents[id]);
+}
+
 void hidScanInput()
 {
 	kOld = kHeld;
@@ -93,7 +106,7 @@ void hidCircleRead(circlePosition* pos)
 	if (pos) *pos = cPos;
 }
 
-Result HIDUSER_GetSharedMem(Handle* outMemHandle)
+Result HIDUSER_GetHandles(Handle* outMemHandle, Handle *eventpad0, Handle *eventpad1, Handle *eventaccel, Handle *eventgyro, Handle *eventdebugpad)
 {
 	u32* cmdbuf=getThreadCommandBuffer();
 	cmdbuf[0]=0xa0000; //request header code
@@ -103,6 +116,12 @@ Result HIDUSER_GetSharedMem(Handle* outMemHandle)
 
 	if(outMemHandle)*outMemHandle=cmdbuf[3];
 
+	if(eventpad0)*eventpad0=cmdbuf[4];
+	if(eventpad1)*eventpad1=cmdbuf[5];
+	if(eventaccel)*eventaccel=cmdbuf[6];
+	if(eventgyro)*eventgyro=cmdbuf[7];
+	if(eventdebugpad)*eventdebugpad=cmdbuf[8];
+
 	return cmdbuf[1];
 }
 

From 62f26e87604f7f351645586d582830860241a140 Mon Sep 17 00:00:00 2001
From: yellows8 <yellows8@users.noreply.github.com>
Date: Sat, 30 Aug 2014 20:41:09 -0400
Subject: [PATCH 02/14] Improved HID code, and added Accelerometer/Gyroscope
 support.

---
 libctru/include/3ds/services/hid.h | 17 +++++++
 libctru/source/services/hid.c      | 74 ++++++++++++++++++++++++++----
 2 files changed, 82 insertions(+), 9 deletions(-)

diff --git a/libctru/include/3ds/services/hid.h b/libctru/include/3ds/services/hid.h
index b39898f..3333abb 100644
--- a/libctru/include/3ds/services/hid.h
+++ b/libctru/include/3ds/services/hid.h
@@ -41,6 +41,20 @@ typedef struct
 	s16 dx, dy;
 } circlePosition;
 
+typedef struct
+{
+	s16 x;
+	s16 y;
+	s16 z;
+} accelVector;
+
+typedef struct
+{
+	s16 x;//roll
+	s16 z;//yaw
+	s16 y;//pitch
+} angularRate;
+
 typedef enum
 {
 	HIDEVENT_PAD0 = 0, //"Event signaled by HID-module, when the sharedmem+0(PAD/circle-pad)/+0xA8(touch-screen) region was updated."
@@ -64,6 +78,8 @@ u32 hidKeysDown();
 u32 hidKeysUp();
 void hidTouchRead(touchPosition* pos);
 void hidCircleRead(circlePosition* pos);
+void hidAccelRead(accelVector* vector);
+void hidGyroRead(angularRate* rate);
 
 void hidWaitForEvent(HID_Event id, bool nextEvent);
 
@@ -80,3 +96,4 @@ Result HIDUSER_EnableAccelerometer();
 Result HIDUSER_DisableAccelerometer();
 Result HIDUSER_EnableGyroscope();
 Result HIDUSER_DisableGyroscope();
+
diff --git a/libctru/source/services/hid.c b/libctru/source/services/hid.c
index aaff8e1..9e32566 100644
--- a/libctru/source/services/hid.c
+++ b/libctru/source/services/hid.c
@@ -15,6 +15,8 @@ vu32* hidSharedMem;
 static u32 kOld, kHeld, kDown, kUp;
 static touchPosition tPos;
 static circlePosition cPos;
+static accelVector aVec;
+static angularRate gRate;
 
 
 Result hidInit(u32* sharedMem)
@@ -62,23 +64,66 @@ void hidWaitForEvent(HID_Event id, bool nextEvent)
 		svcClearEvent(hidEvents[id]);
 }
 
+u32 hidCheckSectionUpdateTime(vu32 *sharedmem_section, u32 id)
+{
+	s64 tick0=0, tick1=0;
+
+	if(id==0)
+	{
+		tick0 = *((u64*)&sharedmem_section[0]);
+		tick1 = *((u64*)&sharedmem_section[2]);
+
+		if(tick0==tick1 || tick0<0 || tick1<0)return 1;
+	}
+
+	return 0;
+}
+
 void hidScanInput()
 {
+	u32 Id=0;
+
 	kOld = kHeld;
 
-	u32 padId = hidSharedMem[4];
-	if(padId>7)padId=7;
-	kHeld = hidSharedMem[10 + padId*4];
-	cPos = *(circlePosition*)&hidSharedMem[10 + padId*4 + 3];
+	kHeld = 0;
+	memset(&cPos, 0, sizeof(circlePosition));
+	memset(&tPos, 0, sizeof(touchPosition));
+	memset(&aVec, 0, sizeof(accelVector));
+	memset(&gRate, 0, sizeof(angularRate));
 
-	u32 touchId = hidSharedMem[42 + 4];
-	if(touchId>7)touchId=7;
-	tPos = *(touchPosition*)&hidSharedMem[42 + 8 + touchId*2];
-	if (hidSharedMem[42 + 8 + touchId*2 + 1])
-		kHeld |= KEY_TOUCH;
+	Id = hidSharedMem[4];//PAD / circle-pad
+	if(Id>7)Id=7;
+	if(hidCheckSectionUpdateTime(hidSharedMem, Id)==0)
+	{
+		kHeld = hidSharedMem[10 + Id*4];
+		cPos = *(circlePosition*)&hidSharedMem[10 + Id*4 + 3];
+	}
+
+	Id = hidSharedMem[42 + 4];//Touch-screen
+	if(Id>7)Id=7;
+	if(hidCheckSectionUpdateTime(&hidSharedMem[42], Id)==0)
+	{
+		tPos = *(touchPosition*)&hidSharedMem[42 + 8 + Id*2];
+		if (hidSharedMem[42 + 8 + Id*2 + 1])
+			kHeld |= KEY_TOUCH;
+	}
 
 	kDown = (~kOld) & kHeld;
 	kUp = kOld & (~kHeld);
+
+	Id = hidSharedMem[66 + 4];//Accelerometer
+	if(Id>7)Id=7;
+	if(hidCheckSectionUpdateTime(&hidSharedMem[66], Id)==0)
+	{
+		aVec = *(accelVector*)&hidSharedMem[66 + 8 + Id*2];
+	}
+
+	Id = hidSharedMem[86 + 4];//Gyroscope
+	if(Id>31)Id=31;
+	if(hidCheckSectionUpdateTime(&hidSharedMem[86], Id)==0)
+	{
+		gRate = *(angularRate*)&hidSharedMem[86 + 8 + Id*2];
+	}
 }
 
 u32 hidKeysHeld()
@@ -106,6 +151,16 @@ void hidCircleRead(circlePosition* pos)
 	if (pos) *pos = cPos;
 }
 
+void hidAccelRead(accelVector* vector)
+{
+	if (vector) *vector = aVec;
+}
+
+void hidGyroRead(angularRate* rate)
+{
+	if (rate) *rate = gRate;
+}
+
 Result HIDUSER_GetHandles(Handle* outMemHandle, Handle *eventpad0, Handle *eventpad1, Handle *eventaccel, Handle *eventgyro, Handle *eventdebugpad)
 {
 	u32* cmdbuf=getThreadCommandBuffer();
@@ -168,3 +223,4 @@ Result HIDUSER_DisableGyroscope()
 
 	return cmdbuf[1];
 }
+

From 3d34e123eea9bf6d8636ad3a719e92bd43691d00 Mon Sep 17 00:00:00 2001
From: profi200 <fd3194@gmx.de>
Date: Mon, 1 Sep 2014 15:49:32 +0200
Subject: [PATCH 03/14] Added functions for sysCore usage

---
 libctru/include/3ds/services/apt.h |  2 ++
 libctru/source/services/apt.c      | 32 ++++++++++++++++++++++++++++++
 2 files changed, 34 insertions(+)

diff --git a/libctru/include/3ds/services/apt.h b/libctru/include/3ds/services/apt.h
index 1a1e644..9377c0b 100644
--- a/libctru/include/3ds/services/apt.h
+++ b/libctru/include/3ds/services/apt.h
@@ -67,3 +67,5 @@ Result APT_ReplySleepQuery(Handle* handle, NS_APPID appID, u32 a);
 Result APT_ReplySleepNotificationComplete(Handle* handle, NS_APPID appID);
 Result APT_PrepareToCloseApplication(Handle* handle, u8 a);
 Result APT_CloseApplication(Handle* handle, u32 a, u32 b, u32 c);
+Result APT_SetAppCpuTimeLimit(Handle* handle, u32 percent);
+Result APT_GetAppCpuTimeLimit(Handle* handle, u32 *percent);
diff --git a/libctru/source/services/apt.c b/libctru/source/services/apt.c
index bda13b6..d2f65d9 100644
--- a/libctru/source/services/apt.c
+++ b/libctru/source/services/apt.c
@@ -743,3 +743,35 @@ Result APT_CloseApplication(Handle* handle, u32 a, u32 b, u32 c)
 
 	return cmdbuf[1];
 }
+
+//See http://3dbrew.org/APT:SetApplicationCpuTimeLimit
+Result APT_SetAppCpuTimeLimit(Handle* handle, u32 percent)
+{
+	if(!handle)handle=&aptuHandle;
+
+	u32* cmdbuf=getThreadCommandBuffer();
+	cmdbuf[0]=0x4F0080;
+	cmdbuf[1]=1;
+	cmdbuf[2]=percent;
+	
+	Result ret=0;
+	if((ret=svcSendSyncRequest(*handle)))return ret;
+
+	return cmdbuf[1];
+}
+
+Result APT_GetAppCpuTimeLimit(Handle* handle, u32 *percent)
+{
+	if(!handle)handle=&aptuHandle;
+
+	u32* cmdbuf=getThreadCommandBuffer();
+	cmdbuf[0]=0x500040;
+	cmdbuf[1]=1;
+	
+	Result ret=0;
+	if((ret=svcSendSyncRequest(*handle)))return ret;
+
+	if(percent)*percent=cmdbuf[2];
+
+	return cmdbuf[1];
+}

From a2f32f91c1ad914ffb50f988288309210c895e19 Mon Sep 17 00:00:00 2001
From: profi200 <fd3194@gmx.de>
Date: Mon, 1 Sep 2014 15:51:53 +0200
Subject: [PATCH 04/14] Now the same as in master

---
 README.md | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/README.md b/README.md
index 01e0c2a..eabf74c 100644
--- a/README.md
+++ b/README.md
@@ -6,3 +6,8 @@ library for writing user mode arm11 code for the 3DS (CTR)
 
 the goal with this is to create a very straightforward interface with the 3DS's OS.
 it is not meant to provide higher level functions; to put things in perspective, the purpose of ctrulib would be to sit between the OS and a possible port of SDL rather than replace it.
+
+license
+=======
+
+you may do whatever the hell you want with ctrulib as long as you give proper credit to its authors. that includes making commercial games/apps that use ctrulib. i'd personally rather you didn't sell ctrulib itself, because that would not be nice of you, but i'm not a lawyer and it's not like i'm going to sue you.

From aa77f9b1e2b42ad79bfbb9f555a17b144680da71 Mon Sep 17 00:00:00 2001
From: yellows8 <yellows8@users.noreply.github.com>
Date: Wed, 3 Sep 2014 14:36:05 -0400
Subject: [PATCH 05/14] Added code for using the microphone and an example app
 for it.

---
 examples/mic/Makefile              | 139 +++++++++++++++
 examples/mic/README.md             |   5 +
 examples/mic/source/main.c         |  95 ++++++++++
 libctru/include/3ds.h              |   1 +
 libctru/include/3ds/services/mic.h |  20 +++
 libctru/source/services/mic.c      | 270 +++++++++++++++++++++++++++++
 6 files changed, 530 insertions(+)
 create mode 100644 examples/mic/Makefile
 create mode 100644 examples/mic/README.md
 create mode 100644 examples/mic/source/main.c
 create mode 100644 libctru/include/3ds/services/mic.h
 create mode 100644 libctru/source/services/mic.c

diff --git a/examples/mic/Makefile b/examples/mic/Makefile
new file mode 100644
index 0000000..9a74360
--- /dev/null
+++ b/examples/mic/Makefile
@@ -0,0 +1,139 @@
+#---------------------------------------------------------------------------------
+.SUFFIXES:
+#---------------------------------------------------------------------------------
+
+ifeq ($(strip $(DEVKITARM)),)
+$(error "Please set DEVKITARM in your environment. export DEVKITARM=<path to>devkitARM")
+endif
+
+ifeq ($(strip $(CTRULIB)),)
+# THIS IS TEMPORARY - in the future it should be at $(DEVKITPRO)/libctru
+$(error "Please set CTRULIB in your environment. export CTRULIB=<path to>libctru")
+endif
+
+TOPDIR ?= $(CURDIR)
+include $(DEVKITARM)/3ds_rules
+
+#---------------------------------------------------------------------------------
+# TARGET is the name of the output
+# BUILD is the directory where object files & intermediate files will be placed
+# SOURCES is a list of directories containing source code
+# DATA is a list of directories containing data files
+# INCLUDES is a list of directories containing header files
+# SPECS is the directory containing the important build and link files
+#---------------------------------------------------------------------------------
+export TARGET		:=	$(shell basename $(CURDIR))
+BUILD		:=	build
+SOURCES		:=	source
+DATA		:=	data
+INCLUDES	:=	include
+
+
+#---------------------------------------------------------------------------------
+# options for code generation
+#---------------------------------------------------------------------------------
+ARCH	:=	-march=armv6k -mtune=mpcore
+
+CFLAGS	:=	-g -Wall -O2 -mword-relocations -save-temps \
+			-fomit-frame-pointer -ffast-math \
+			$(ARCH)
+
+CFLAGS	+=	$(INCLUDE) -DARM11 -D_3DS
+
+CXXFLAGS	:= $(CFLAGS) -fno-rtti -fno-exceptions -std=gnu++11
+
+ASFLAGS	:=	-g $(ARCH)
+LDFLAGS	=	-specs=3dsx.specs -g $(ARCH) \
+			-Wl,-Map,$(TARGET).map
+
+LIBS	:= -lctru -lm
+
+#---------------------------------------------------------------------------------
+# list of directories containing libraries, this must be the top level containing
+# include and lib
+#---------------------------------------------------------------------------------
+LIBDIRS	:= $(CTRULIB)
+ 
+  
+#---------------------------------------------------------------------------------
+# no real need to edit anything past this point unless you need to add additional
+# rules for different file extensions
+#---------------------------------------------------------------------------------
+ifneq ($(BUILD),$(notdir $(CURDIR)))
+#---------------------------------------------------------------------------------
+ 
+export OUTPUT	:=	$(CURDIR)/$(TARGET)
+export TOPDIR	:=	$(CURDIR)
+
+export VPATH	:=	$(foreach dir,$(SOURCES),$(CURDIR)/$(dir)) \
+			$(foreach dir,$(DATA),$(CURDIR)/$(dir))
+
+export DEPSDIR	:=	$(CURDIR)/$(BUILD)
+
+CFILES		:=	$(foreach dir,$(SOURCES),$(notdir $(wildcard $(dir)/*.c)))
+CPPFILES	:=	$(foreach dir,$(SOURCES),$(notdir $(wildcard $(dir)/*.cpp)))
+SFILES		:=	$(foreach dir,$(SOURCES),$(notdir $(wildcard $(dir)/*.s)))
+BINFILES	:=	$(foreach dir,$(DATA),$(notdir $(wildcard $(dir)/*.*)))
+
+#---------------------------------------------------------------------------------
+# use CXX for linking C++ projects, CC for standard C
+#---------------------------------------------------------------------------------
+ifeq ($(strip $(CPPFILES)),)
+#---------------------------------------------------------------------------------
+	export LD	:=	$(CC)
+#---------------------------------------------------------------------------------
+else
+#---------------------------------------------------------------------------------
+	export LD	:=	$(CXX)
+#---------------------------------------------------------------------------------
+endif
+#---------------------------------------------------------------------------------
+
+export OFILES	:=	$(addsuffix .o,$(BINFILES)) \
+			$(CPPFILES:.cpp=.o) $(CFILES:.c=.o) $(SFILES:.s=.o)
+
+export INCLUDE	:=	$(foreach dir,$(INCLUDES),-I$(CURDIR)/$(dir)) \
+			$(foreach dir,$(LIBDIRS),-I$(dir)/include) \
+			-I$(CURDIR)/$(BUILD)
+
+export LIBPATHS	:=	$(foreach dir,$(LIBDIRS),-L$(dir)/lib)
+
+.PHONY: $(BUILD) clean all
+ 
+#---------------------------------------------------------------------------------
+all: $(BUILD)
+
+$(BUILD):
+	@[ -d $@ ] || mkdir -p $@
+	@make --no-print-directory -C $(BUILD) -f $(CURDIR)/Makefile
+ 
+#---------------------------------------------------------------------------------
+clean:
+	@echo clean ...
+	@rm -fr $(BUILD) $(TARGET).3dsx $(TARGET).elf
+ 
+ 
+#---------------------------------------------------------------------------------
+else
+ 
+DEPENDS	:=	$(OFILES:.o=.d)
+ 
+#---------------------------------------------------------------------------------
+# main targets
+#---------------------------------------------------------------------------------
+$(OUTPUT).3dsx	:	$(OUTPUT).elf
+$(OUTPUT).elf	:	$(OFILES)
+
+#---------------------------------------------------------------------------------
+# you need a rule like this for each extension you use as binary data 
+#---------------------------------------------------------------------------------
+%.bin.o	:	%.bin
+#---------------------------------------------------------------------------------
+	@echo $(notdir $<)
+	$(bin2o)
+
+-include $(DEPENDS)
+ 
+#---------------------------------------------------------------------------------------
+endif
+#---------------------------------------------------------------------------------------
diff --git a/examples/mic/README.md b/examples/mic/README.md
new file mode 100644
index 0000000..37a7328
--- /dev/null
+++ b/examples/mic/README.md
@@ -0,0 +1,5 @@
+arm11u
+=======
+
+Example for using the microphone with ctrulib. Hold down the A button to record, the app will then play the recorded audio once the A button is released. Roughly 32 seconds of audio can be recorded with the default audiobuf size in this app.
+
diff --git a/examples/mic/source/main.c b/examples/mic/source/main.c
new file mode 100644
index 0000000..022902a
--- /dev/null
+++ b/examples/mic/source/main.c
@@ -0,0 +1,95 @@
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <malloc.h>
+#include <3ds.h>
+
+int main()
+{
+	u8 *framebuf;
+	u32 *sharedmem = NULL, sharedmem_size = 0x30000;
+	u8 *audiobuf;
+	u32 audiobuf_size = 0x100000, audiobuf_pos = 0;
+	u8 control=0x40;
+
+	srvInit();	
+	aptInit();
+	gfxInit();
+	hidInit(NULL);
+	aptSetupEventHandler();
+	
+	CSND_initialize(NULL);
+
+	sharedmem = (u32*)memalign(0x1000, sharedmem_size);
+	audiobuf = linearAlloc(audiobuf_size);
+
+	MIC_Initialize(sharedmem, sharedmem_size, control, 0, 3, 1, 1);//See mic.h.
+
+	APP_STATUS status;
+	while((status=aptGetStatus())!=APP_EXITING)
+	{
+		if(status==APP_RUNNING)
+		{
+			framebuf = gfxGetFramebuffer(GFX_TOP, GFX_LEFT, NULL, NULL);
+			hidScanInput();
+
+			if(hidKeysDown() & KEY_A)
+			{
+				audiobuf_pos = 0;
+
+				CSND_setchannel_playbackstate(0x8, 0);//Stop audio playback.
+				CSND_sharedmemtype0_cmdupdatestate(0);
+
+				MIC_SetRecording(1);
+
+				memset(framebuf, 0x20, 0x46500);
+			}
+
+			if((hidKeysHeld() & KEY_A) && audiobuf_pos < audiobuf_size)
+			{
+				audiobuf_pos+= MIC_ReadAudioData(&audiobuf[audiobuf_pos], audiobuf_size-audiobuf_pos, 1);
+				if(audiobuf_pos > audiobuf_size)audiobuf_pos = audiobuf_size;
+
+				memset(framebuf, 0x60, 0x46500);
+			}
+
+			if(hidKeysUp() & KEY_A)
+			{
+				MIC_SetRecording(0);
+				GSPGPU_FlushDataCache(NULL, audiobuf, audiobuf_pos);
+				CSND_playsound(0x8, CSND_LOOP_DISABLE, CSND_ENCODING_PCM16, 16000, (u32*)audiobuf, NULL, audiobuf_pos, 2, 0);
+
+				memset(framebuf, 0xe0, 0x46500);
+
+				gfxFlushBuffers();
+				gfxSwapBuffers();
+
+				framebuf = gfxGetFramebuffer(GFX_TOP, GFX_LEFT, NULL, NULL);
+				memset(framebuf, 0xe0, 0x46500);
+			}
+
+			gfxFlushBuffers();
+			gfxSwapBuffers();
+		}
+		else if(status == APP_SUSPENDING)
+		{
+			aptReturnToMenu();
+		}
+		else if(status == APP_SLEEPMODE)
+		{
+			aptWaitStatusEvent();
+		}
+		gspWaitForVBlank();
+	}
+
+	MIC_Shutdown();
+
+	CSND_shutdown();
+
+	hidExit();
+	gfxExit();
+	aptExit();
+	srvExit();
+	return 0;
+}
+
diff --git a/libctru/include/3ds.h b/libctru/include/3ds.h
index f2a9367..1f17a92 100644
--- a/libctru/include/3ds.h
+++ b/libctru/include/3ds.h
@@ -23,6 +23,7 @@ extern "C" {
 #include <3ds/services/ir.h>
 #include <3ds/services/ptm.h>
 #include <3ds/services/soc.h>
+#include <3ds/services/mic.h>
 
 #include <3ds/gpu/gx.h>
 #include <3ds/gpu/gpu.h>
diff --git a/libctru/include/3ds/services/mic.h b/libctru/include/3ds/services/mic.h
new file mode 100644
index 0000000..e0c0cae
--- /dev/null
+++ b/libctru/include/3ds/services/mic.h
@@ -0,0 +1,20 @@
+#pragma once
+
+//See also: http://3dbrew.org/wiki/MIC_Services
+
+Result MIC_Initialize(u32 *sharedmem, u32 sharedmem_size, u8 control, u8 recording, u8 unk0, u8 unk1, u8 unk2);//sharedmem_size = audiodata size + 4, aligned to 0x1000-bytes. The sharedmem ptr must be 0x1000-bytes aligned. The offical 3ds-sound app uses the following values for unk0-unk2: 3, 1, and 1.
+Result MIC_Shutdown();
+u32 MIC_GetSharedMemOffsetValue();
+u32 MIC_ReadAudioData(u8 *outbuf, u32 readsize, u32 waitforevent);//Reads MIC audio data. When waitforevent is non-zero, this clears the event, then waits for MIC-module to signal it again when audio data is written to shared-mem. The return value is the actual byte-size of the read data.
+
+Result MIC_MapSharedMem(Handle handle, u32 size);
+Result MIC_UnmapSharedMem();
+Result MIC_cmd3_Initialize(u8 unk0, u8 unk1, u32 sharedmem_baseoffset, u32 sharedmem_endoffset, u8 unk2);
+Result MIC_cmd5();
+Result MIC_GetCNTBit15(u8 *out);
+Result MIC_GetEventHandle(Handle *handle);
+Result MIC_SetControl(u8 value);//See here: http://3dbrew.org/wiki/MIC_Services
+Result MIC_GetControl(u8 *value);
+Result MIC_SetRecording(u8 value);
+Result MIC_IsRecoding(u8 *value);
+
diff --git a/libctru/source/services/mic.c b/libctru/source/services/mic.c
new file mode 100644
index 0000000..395792b
--- /dev/null
+++ b/libctru/source/services/mic.c
@@ -0,0 +1,270 @@
+#include <stdlib.h>
+#include <string.h>
+#include <3ds.h>
+
+//See also: http://3dbrew.org/wiki/MIC_Services
+
+Handle MIC_handle;
+
+static u8 *MIC_sharedmem;
+static u32 MIC_sharedmem_size;
+static u32 *MIC_sharedmem_offsetfield, MIC_sharedmem_offsetfield_location;
+static Handle MIC_sharedmem_handle;
+static Handle MIC_event;
+
+static u32 MIC_prev_endpos, MIC_cur_endpos;
+
+Result MIC_Initialize(u32 *sharedmem, u32 sharedmem_size, u8 control, u8 recording, u8 unk0, u8 unk1, u8 unk2)
+{
+	Result ret=0;
+
+	MIC_sharedmem = (u8*)sharedmem;
+	MIC_sharedmem_size = sharedmem_size;
+	MIC_sharedmem_offsetfield_location = MIC_sharedmem_size - 4;
+	MIC_sharedmem_offsetfield = (u32*)&MIC_sharedmem[MIC_sharedmem_offsetfield_location];
+	MIC_event = 0;
+	MIC_prev_endpos = 0;
+	MIC_cur_endpos = 0;
+
+	ret = srvGetServiceHandle(&MIC_handle, "mic:u");
+	if(ret!=0)return ret;
+
+	ret = svcCreateMemoryBlock(&MIC_sharedmem_handle, (u32)MIC_sharedmem, MIC_sharedmem_size, 3, 3);
+	if(ret!=0)return ret;
+
+	ret = MIC_SetControl(control);
+	if(ret!=0)return ret;
+
+	ret = MIC_MapSharedMem(MIC_sharedmem_handle, sharedmem_size);
+	if(ret!=0)return ret;
+
+	ret = MIC_SetRecording(recording);
+	if(ret!=0)return ret;
+
+	ret = MIC_cmd3_Initialize(unk0, unk1, 0, MIC_sharedmem_size-4, unk2);
+	if(ret!=0)return ret;
+
+	ret = MIC_GetEventHandle(&MIC_event);
+	if(ret!=0)return ret;
+	svcClearEvent(MIC_event);
+
+	return 0;
+}
+
+Result MIC_Shutdown()
+{
+	Result ret=0;
+
+	MIC_cmd5();
+	MIC_SetRecording(0);
+
+	ret = MIC_UnmapSharedMem();
+	if(ret!=0)return ret;
+
+	MIC_cmd5();
+
+	ret = svcUnmapMemoryBlock(MIC_sharedmem_handle, (u32)MIC_sharedmem);
+	if(ret!=0)return ret;
+
+	ret = svcCloseHandle(MIC_sharedmem_handle);
+	if(ret!=0)return ret;
+
+	ret = svcCloseHandle(MIC_event);
+	if(ret!=0)return ret;
+
+	ret = svcCloseHandle(MIC_handle);
+	if(ret!=0)return ret;
+
+	MIC_sharedmem_offsetfield = NULL;
+	MIC_sharedmem = NULL;
+	MIC_sharedmem_size = 0;
+	MIC_handle = 0;
+	MIC_event = 0;
+
+	return 0;
+}
+
+u32 MIC_GetSharedMemOffsetValue()
+{
+	u32 pos = 0;
+
+	if(MIC_sharedmem_offsetfield==NULL)return pos;
+	pos = *MIC_sharedmem_offsetfield;
+	if(pos > MIC_sharedmem_offsetfield_location)pos = MIC_sharedmem_offsetfield_location;
+
+	return pos;
+}
+
+u32 MIC_ReadAudioData(u8 *outbuf, u32 readsize, u32 waitforevent)
+{
+	u32 pos = 0, bufpos = 0;
+	
+	if(waitforevent)
+	{
+		svcClearEvent(MIC_event);
+		svcWaitSynchronization(MIC_event, U64_MAX);
+	}
+
+	MIC_prev_endpos = MIC_cur_endpos;
+	MIC_cur_endpos = MIC_GetSharedMemOffsetValue();
+	pos = MIC_prev_endpos;
+
+	while(pos != MIC_cur_endpos)
+	{
+		if(pos >= MIC_sharedmem_offsetfield_location)pos = 0;
+		if(bufpos>=readsize)break;
+
+		outbuf[bufpos] = MIC_sharedmem[pos];
+		bufpos++;
+		pos++;
+	}
+
+	return bufpos;
+}
+
+Result MIC_MapSharedMem(Handle handle, u32 size)
+{
+	Result ret=0;
+	u32 *cmdbuf = getThreadCommandBuffer();
+
+	cmdbuf[0] = 0x00010042;
+	cmdbuf[1] = size;
+	cmdbuf[2] = 0;
+	cmdbuf[3] = handle;
+
+	if((ret = svcSendSyncRequest(MIC_handle))!=0)return ret;
+
+	return (Result)cmdbuf[1];
+}
+
+Result MIC_UnmapSharedMem()
+{
+	Result ret=0;
+	u32 *cmdbuf = getThreadCommandBuffer();
+
+	cmdbuf[0] = 0x00020000;
+
+	if((ret = svcSendSyncRequest(MIC_handle))!=0)return ret;
+
+	return (Result)cmdbuf[1];
+}
+
+Result MIC_cmd3_Initialize(u8 unk0, u8 unk1, u32 sharedmem_baseoffset, u32 sharedmem_endoffset, u8 unk2)
+{
+	Result ret=0;
+	u32 *cmdbuf = getThreadCommandBuffer();
+
+	cmdbuf[0] = 0x00030140;
+	cmdbuf[1] = unk0;
+	cmdbuf[2] = unk1;
+	cmdbuf[3] = sharedmem_baseoffset;
+	cmdbuf[4] = sharedmem_endoffset;
+	cmdbuf[5] = unk2;
+
+	if((ret = svcSendSyncRequest(MIC_handle))!=0)return ret;
+
+	return (Result)cmdbuf[1];
+}
+
+Result MIC_cmd5()
+{
+	Result ret=0;
+	u32 *cmdbuf = getThreadCommandBuffer();
+
+	cmdbuf[0] = 0x00050000;
+
+	if((ret = svcSendSyncRequest(MIC_handle))!=0)return ret;
+
+	return (Result)cmdbuf[1];
+}
+
+Result MIC_GetCNTBit15(u8 *out)
+{
+	Result ret=0;
+	u32 *cmdbuf = getThreadCommandBuffer();
+
+	cmdbuf[0] = 0x00060000;
+
+	if((ret = svcSendSyncRequest(MIC_handle))!=0)return ret;
+
+	if(out)*out = cmdbuf[2];
+
+	return (Result)cmdbuf[1];
+}
+
+Result MIC_GetEventHandle(Handle *handle)
+{
+	Result ret=0;
+	u32 *cmdbuf = getThreadCommandBuffer();
+
+	if(MIC_event)
+	{
+		*handle = MIC_event;
+		return 0;
+	}
+
+	cmdbuf[0] = 0x00070000;
+
+	if((ret = svcSendSyncRequest(MIC_handle))!=0)return ret;
+
+	if(handle)*handle = cmdbuf[2];
+
+	return (Result)cmdbuf[1];
+}
+
+Result MIC_SetControl(u8 value)
+{
+	Result ret=0;
+	u32 *cmdbuf = getThreadCommandBuffer();
+
+	cmdbuf[0] = 0x00080040;
+	cmdbuf[1] = value;
+
+	if((ret = svcSendSyncRequest(MIC_handle))!=0)return ret;
+
+	return (Result)cmdbuf[1];
+}
+
+Result MIC_GetControl(u8 *value)
+{
+	Result ret=0;
+	u32 *cmdbuf = getThreadCommandBuffer();
+
+	cmdbuf[0] = 0x00090000;
+
+	if((ret = svcSendSyncRequest(MIC_handle))!=0)return ret;
+
+	if(value)*value = cmdbuf[2];
+
+	return (Result)cmdbuf[1];
+}
+
+Result MIC_SetRecording(u8 value)
+{
+	Result ret=0;
+	u32 *cmdbuf = getThreadCommandBuffer();
+
+	cmdbuf[0] = 0x000A0040;
+	cmdbuf[1] = value;
+
+	if((ret = svcSendSyncRequest(MIC_handle))!=0)return ret;
+
+	if(value==1)MIC_cur_endpos = MIC_GetSharedMemOffsetValue();
+
+	return (Result)cmdbuf[1];
+}
+
+Result MIC_IsRecoding(u8 *value)
+{
+	Result ret=0;
+	u32 *cmdbuf = getThreadCommandBuffer();
+
+	cmdbuf[0] = 0x000B0000;
+
+	if((ret = svcSendSyncRequest(MIC_handle))!=0)return ret;
+
+	if(value)*value = cmdbuf[2];
+
+	return (Result)cmdbuf[1];
+}
+

From 63c3e5af0d9d800937a77308b0479b01b7947d78 Mon Sep 17 00:00:00 2001
From: yellows8 <yellows8@users.noreply.github.com>
Date: Wed, 3 Sep 2014 14:38:47 -0400
Subject: [PATCH 06/14] Fixed mic example README.

---
 examples/mic/README.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/examples/mic/README.md b/examples/mic/README.md
index 37a7328..f3e6d72 100644
--- a/examples/mic/README.md
+++ b/examples/mic/README.md
@@ -1,4 +1,4 @@
-arm11u
+mic
 =======
 
 Example for using the microphone with ctrulib. Hold down the A button to record, the app will then play the recorded audio once the A button is released. Roughly 32 seconds of audio can be recorded with the default audiobuf size in this app.

From 449ead1141466c59bb48c242520d58eb8567b318 Mon Sep 17 00:00:00 2001
From: plutoo <plutoo@univor.se>
Date: Sat, 6 Sep 2014 21:07:26 +0200
Subject: [PATCH 07/14] os: Added osGetTime().

---
 libctru/include/3ds/os.h |  1 +
 libctru/source/os.c      | 31 +++++++++++++++++++++++++++++++
 2 files changed, 32 insertions(+)

diff --git a/libctru/include/3ds/os.h b/libctru/include/3ds/os.h
index c267590..92e6483 100644
--- a/libctru/include/3ds/os.h
+++ b/libctru/include/3ds/os.h
@@ -7,3 +7,4 @@ u32 osConvertVirtToPhys(u32 vaddr);
 const char* osStrError(u32 error);
 u32 osGetFirmVersion();
 u32 osGetKernelVersion();
+u64 osGetTime();
diff --git a/libctru/source/os.c b/libctru/source/os.c
index f09bf24..84431bf 100644
--- a/libctru/source/os.c
+++ b/libctru/source/os.c
@@ -1,5 +1,21 @@
 #include <3ds.h>
 
+#define TICKS_PER_MSEC 268123.480
+
+typedef struct {
+	u64 date_time;
+	u64 update_tick;
+	//...
+} datetime_t;
+
+static volatile u32* __datetime_selector =
+	(u32*) 0x1FF81000;
+static volatile datetime_t* __datetime1 =
+	(datetime_t*) 0x1FF81020;
+static volatile datetime_t* __datetime2 =
+	(datetime_t*) 0x1FF81040;
+
+
 u32 osConvertVirtToPhys(u32 vaddr)
 {
 	if(vaddr >= 0x14000000 && vaddr < 0x1c000000)
@@ -13,6 +29,21 @@ u32 osConvertVirtToPhys(u32 vaddr)
 	return 0;
 }
 
+// Returns number of milliseconds since 1st Jan 1900 00:00.
+u64 osGetTime() {
+	volatile datetime_t* dt;
+
+	switch(*__datetime_selector & 1) {
+	case 0:
+		dt = __datetime1;
+	case 1:
+		dt = __datetime2;
+	}
+
+	u64 offset = (svcGetSystemTick() - dt->update_tick) / TICKS_PER_MSEC;
+	return dt->date_time + offset;
+}
+
 u32 osGetFirmVersion() {
 	return (*(u32*)0x1FF80000) & ~0xFF;
 }

From a9f4b3402beea2cf69728bda40572a387b640dda Mon Sep 17 00:00:00 2001
From: yellows8 <yellows8@users.noreply.github.com>
Date: Sat, 6 Sep 2014 17:49:48 -0400
Subject: [PATCH 08/14] Fixed broken CSND_playsound() code for looping(use
 vaddr1 for that instead of vaddr0, and adjust the size).

---
 libctru/source/services/csnd.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/libctru/source/services/csnd.c b/libctru/source/services/csnd.c
index 9eb11bb..77b1686 100644
--- a/libctru/source/services/csnd.c
+++ b/libctru/source/services/csnd.c
@@ -285,7 +285,11 @@ Result CSND_playsound(u32 channel, u32 looping, u32 encoding, u32 samplerate, u3
 
 	CSND_sharedmemtype0_cmde(channel, looping, encoding, samplerate, unk0, unk1, physaddr0, physaddr1, totalbytesize);
 	CSND_sharedmemtype0_cmd8(channel, samplerate);
-	if(looping)CSND_sharedmemtype0_cmd3(channel, physaddr0, totalbytesize);
+	if(looping)
+	{
+		if(physaddr1>physaddr0)totalbytesize-= (u32)physaddr1 - (u32)physaddr0;
+		CSND_sharedmemtype0_cmd3(channel, physaddr1, totalbytesize);
+	}
 	CSND_sharedmemtype0_cmd8(channel, samplerate);
 	CSND_sharedmemtype0_cmd9(channel, 0xffff);
 	CSND_setchannel_playbackstate(channel, 1);

From ad08977e3777de77fe12c7124e2148f1a78c101a Mon Sep 17 00:00:00 2001
From: StapleButter <thetotalworm@gmail.com>
Date: Thu, 11 Sep 2014 21:02:56 +0200
Subject: [PATCH 09/14] Fix FS init.

---
 libctru/source/services/fs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libctru/source/services/fs.c b/libctru/source/services/fs.c
index be5bbd0..00a184b 100644
--- a/libctru/source/services/fs.c
+++ b/libctru/source/services/fs.c
@@ -68,7 +68,7 @@ FSUSER_Initialize(Handle* handle)
 	{
 		// don't run command if we got handle from the list
 		handle = &fsuHandle;
-		if(fsuHandle != 0 && __get_handle_from_list("fs:USER")==0)
+		if(fsuHandle != 0 && __get_handle_from_list("fs:USER")!=0)
 			return 0;
 	}
 

From 8b27dbe623c2d38b447bdfa9c2b16c0ba5183887 Mon Sep 17 00:00:00 2001
From: StapleButter <thetotalworm@gmail.com>
Date: Sat, 13 Sep 2014 17:01:20 +0200
Subject: [PATCH 10/14] * Make the GSP event handler signal events in the right
 order (essential for proper PICA200 sync) * Add some comments about the GSP
 events (based on my observations, may not be right)

---
 libctru/include/3ds/services/gsp.h | 6 +++---
 libctru/source/services/gsp.c      | 8 ++++----
 2 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/libctru/include/3ds/services/gsp.h b/libctru/include/3ds/services/gsp.h
index f7ddfc2..d324b9a 100644
--- a/libctru/include/3ds/services/gsp.h
+++ b/libctru/include/3ds/services/gsp.h
@@ -37,12 +37,12 @@ typedef struct
 
 typedef enum
 {
-	GSPEVENT_PSC0 = 0,
+	GSPEVENT_PSC0 = 0,	// memory fill completed
 	GSPEVENT_PSC1,
 	GSPEVENT_VBlank0,
 	GSPEVENT_VBlank1,
-	GSPEVENT_PPF,
-	GSPEVENT_P3D,
+	GSPEVENT_PPF,		// display transfer finished
+	GSPEVENT_P3D,		// command list processing finished
 	GSPEVENT_DMA,
 
 	GSPEVENT_MAX, // used to know how many events there are
diff --git a/libctru/source/services/gsp.c b/libctru/source/services/gsp.c
index 7c64019..24fff1d 100644
--- a/libctru/source/services/gsp.c
+++ b/libctru/source/services/gsp.c
@@ -85,15 +85,15 @@ void gspEventThreadMain(u32 arg)
 		svcClearEvent(gspEvent);
 
 		int count = gspEventData[1];
-		int last = gspEventData[0] + count;
+		int cur = gspEventData[0];
+		int last = cur + count;
 		while (last >= 0x34) last -= 0x34;
-		int cur = last;
 		int i;
 		for (i = 0; i < count; i ++)
 		{
 			int curEvt = gspEventData[0xC + cur];
-			cur --;
-			if (cur < 0) cur += 0x34;
+			cur ++;
+			if (cur >= 0x34) cur -= 0x34;
 			if (curEvt >= GSPEVENT_MAX) continue;
 			svcSignalEvent(gspEvents[curEvt]);
 		}

From 2c589cd2ef9a2b663e7c792e0451af1f1d544223 Mon Sep 17 00:00:00 2001
From: fincs <fincs.alt1@gmail.com>
Date: Tue, 16 Sep 2014 22:24:26 +0200
Subject: [PATCH 11/14] linearAlloc: use 16-byte alignment

---
 libctru/source/linear.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/libctru/source/linear.c b/libctru/source/linear.c
index ba78f40..66eb43c 100644
--- a/libctru/source/linear.c
+++ b/libctru/source/linear.c
@@ -5,20 +5,20 @@ extern u32 __linear_heap, __linear_heap_size;
 // TODO: this allocator sucks! It is not thread-safe and you cannot 'free' this memory.
 void* linearAlloc(size_t size)
 {
-    static size_t currentOffset = 0;
-    size_t free = __linear_heap_size - currentOffset;
+	static size_t currentOffset = 0;
+	size_t free = __linear_heap_size - currentOffset;
 
-	// Enforce 8-byte alignment
-	size = (size + 7) &~ 7;
+	// Enforce 16-byte alignment
+	size = (size + 15) &~ 15;
 
 	void* mem = NULL;
-    if (free >= size)
-    {
+	if (free >= size)
+	{
 		mem = (void*)(__linear_heap + currentOffset);
-        currentOffset += size;
-    }
+		currentOffset += size;
+	}
 
-    return mem;
+	return mem;
 }
 
 void* linearRealloc(void* mem, size_t size)

From 27be99a79306526f9da16ebe6bf31fd7f3c5e1ec Mon Sep 17 00:00:00 2001
From: fincs <fincs.alt1@gmail.com>
Date: Wed, 17 Sep 2014 00:22:30 +0200
Subject: [PATCH 12/14] Add proper linear heap allocator

---
 libctru/Makefile                      |   2 +-
 libctru/source/allocator/linear.cpp   |  50 +++++++++++
 libctru/source/allocator/mem_pool.cpp | 125 ++++++++++++++++++++++++++
 libctru/source/allocator/mem_pool.h   |  88 ++++++++++++++++++
 libctru/source/linear.c               |  32 -------
 5 files changed, 264 insertions(+), 33 deletions(-)
 create mode 100644 libctru/source/allocator/linear.cpp
 create mode 100644 libctru/source/allocator/mem_pool.cpp
 create mode 100644 libctru/source/allocator/mem_pool.h
 delete mode 100644 libctru/source/linear.c

diff --git a/libctru/Makefile b/libctru/Makefile
index 83aa6d0..c42536b 100644
--- a/libctru/Makefile
+++ b/libctru/Makefile
@@ -17,7 +17,7 @@ include $(DEVKITARM)/base_rules
 #---------------------------------------------------------------------------------
 TARGET		:=	ctru
 BUILD		:=	build
-SOURCES		:=	source source/services source/gpu
+SOURCES		:=	source source/services source/gpu source/allocator
 DATA		:=	data
 INCLUDES	:=	include
 
diff --git a/libctru/source/allocator/linear.cpp b/libctru/source/allocator/linear.cpp
new file mode 100644
index 0000000..a7581e8
--- /dev/null
+++ b/libctru/source/allocator/linear.cpp
@@ -0,0 +1,50 @@
+#include <3ds.h>
+#include "mem_pool.h"
+
+extern u32 __linear_heap, __linear_heap_size;
+
+static MemPool sLinearPool;
+
+static bool linearInit()
+{
+	auto blk = MemBlock::Create((u8*)__linear_heap, __linear_heap_size);
+	if (blk)
+	{
+		sLinearPool.AddBlock(blk);
+		return true;
+	}
+	return false;
+}
+
+void* linearAlloc(size_t size)
+{
+	// Initialize the pool if it is not ready
+	if (!sLinearPool.Ready() && !linearInit())
+		return nullptr;
+
+	// Reserve memory for MemChunk structure
+	size += 16;
+
+	// Allocate the chunk
+	MemChunk chunk;
+	if (!sLinearPool.Allocate(chunk, size, 4)) // 16-byte alignment
+		return nullptr;
+
+	// Copy the MemChunk structure and return memory
+	auto addr = chunk.addr;
+	*(MemChunk*)addr = chunk;
+	return addr + 16;
+}
+
+void* linearRealloc(void* mem, size_t size)
+{
+	// TODO
+	return NULL;
+}
+
+void linearFree(void* mem)
+{
+	// Find MemChunk structure and free the chunk
+	auto pChunk = (MemChunk*)((u8*)mem - 16);
+	sLinearPool.Deallocate(*pChunk);
+}
diff --git a/libctru/source/allocator/mem_pool.cpp b/libctru/source/allocator/mem_pool.cpp
new file mode 100644
index 0000000..93eeb4f
--- /dev/null
+++ b/libctru/source/allocator/mem_pool.cpp
@@ -0,0 +1,125 @@
+#include "mem_pool.h"
+
+/*
+// This method is currently unused
+void MemPool::CoalesceLeft(MemBlock* b)
+{
+	auto curPtr = b->base;
+	for (auto p = b->prev; p; p = p->prev)
+	{
+		if ((p->base + p->size) != curPtr) break;
+		curPtr = p->base;
+		p->size += b->size;
+		DelBlock(b);
+		b = p;
+	}
+}
+*/
+
+void MemPool::CoalesceRight(MemBlock* b)
+{
+	auto curPtr = b->base + b->size;
+	auto next = b->next;
+	for (auto n = next; n; n = next)
+	{
+		next = n->next;
+		if (n->base != curPtr) break;
+		b->size += n->size;
+		curPtr += n->size;
+		DelBlock(n);
+	}
+}
+
+bool MemPool::Allocate(MemChunk& chunk, u32 size, int align)
+{
+	int alignM = (1 << align) - 1;
+	size = (size + alignM) &~ alignM; // Round the size
+	// Find the first suitable block
+	for (auto b = first; b; b = b->next)
+	{
+		auto addr = b->base;
+		u32 begWaste = (u32)addr & alignM;
+		addr += begWaste;
+		u32 bSize = b->size - begWaste;
+		if (bSize < size) continue;
+
+		// Found space!
+		chunk.addr = addr;
+		chunk.size = size;
+
+		// Resize the block
+		if (!begWaste)
+		{
+			b->base += size;
+			b->size -= size;
+			if (!b->size)
+				DelBlock(b);
+		} else
+		{
+			auto nAddr = addr + size;
+			auto nSize = bSize - size;
+			b->size = begWaste;
+			if (nSize)
+			{
+				// We need to add the tail chunk that wasn't used to the list
+				auto n = MemBlock::Create(nAddr, nSize);
+				if (n) InsertAfter(b, n);
+				else   chunk.size += nSize; // we have no choice but to waste the space.
+			}
+		}
+		return true;
+	}
+
+	return false;
+}
+
+void MemPool::Deallocate(const MemChunk& chunk)
+{
+	u8*  cAddr = chunk.addr;
+	auto cSize = chunk.size;
+	bool done = false;
+
+	// Try to merge the chunk somewhere into the list
+	for (auto b = first; !done && b; b = b->next)
+	{
+		auto addr = b->base;
+		if (addr > cAddr)
+		{
+			if ((cAddr + cSize) == addr)
+			{
+				// Merge the chunk to the left of the block
+				b->base = cAddr;
+				b->size += cSize;
+			} else
+			{
+				// We need to insert a new block
+				auto c = MemBlock::Create(cAddr, cSize);
+				if (c) InsertBefore(b, c);
+			}
+			done = true;
+		} else if ((b->base + b->size) == cAddr)
+		{
+			// Coalesce to the right
+			b->size += cSize;
+			CoalesceRight(b);
+			done = true;
+		}
+	}
+
+	if (!done)
+	{
+		// Either the list is empty or the chunk address is past the end
+		// address of the last block -- let's add a new block at the end
+		auto b = MemBlock::Create(cAddr, cSize);
+		if (b) AddBlock(b);
+	}
+}
+
+/*
+void MemPool::Dump(const char* title)
+{
+	printf("<%s> VRAM Pool Dump\n", title);
+	for (auto b = first; b; b = b->next)
+		printf("  - %p (%u bytes)\n", b->base, b->size);
+}
+*/
diff --git a/libctru/source/allocator/mem_pool.h b/libctru/source/allocator/mem_pool.h
new file mode 100644
index 0000000..7237f49
--- /dev/null
+++ b/libctru/source/allocator/mem_pool.h
@@ -0,0 +1,88 @@
+#pragma once
+#include <3ds.h>
+
+struct MemChunk
+{
+	u8* addr;
+	u32 size;
+};
+
+struct MemBlock
+{
+	MemBlock *prev, *next;
+	u8* base;
+	u32 size;
+
+	static MemBlock* Create(u8* base, u32 size)
+	{
+		auto b = (MemBlock*)malloc(sizeof(MemBlock));
+		if (!b) return nullptr;
+		b->prev = nullptr;
+		b->next = nullptr;
+		b->base = base;
+		b->size = size;
+		return b;
+	}
+};
+
+struct MemPool
+{
+	MemBlock *first, *last;
+
+	bool Ready() { return first != nullptr; }
+
+	void AddBlock(MemBlock* blk)
+	{
+		blk->prev = last;
+		if (last) last->next = blk;
+		if (!first) first = blk;
+		last = blk;
+	}
+
+	void DelBlock(MemBlock* b)
+	{
+		auto prev = b->prev, &pNext = prev ? prev->next : first;
+		auto next = b->next, &nPrev = next ? next->prev : last;
+		pNext = next;
+		nPrev = prev;
+		free(b);
+	}
+
+	void InsertBefore(MemBlock* b, MemBlock* p)
+	{
+		auto prev = b->prev, &pNext = prev ? prev->next : first;
+		b->prev = p;
+		p->next = b;
+		p->prev = prev;
+		pNext = p;
+	}
+
+	void InsertAfter(MemBlock* b, MemBlock* n)
+	{
+		auto next = b->next, &nPrev = next ? next->prev : last;
+		b->next = n;
+		n->prev = b;
+		n->next = next;
+		nPrev = n;
+	}
+
+	//void CoalesceLeft(MemBlock* b);
+	void CoalesceRight(MemBlock* b);
+
+	bool Allocate(MemChunk& chunk, u32 size, int align);
+	void Deallocate(const MemChunk& chunk);
+
+	void Destroy()
+	{
+		MemBlock* next = nullptr;
+		for (auto b = first; b; b = next)
+		{
+			next = b->next;
+			free(b);
+		}
+		first = nullptr;
+		last = nullptr;
+	}
+
+	//void Dump(const char* title);
+};
diff --git a/libctru/source/linear.c b/libctru/source/linear.c
deleted file mode 100644
index 66eb43c..0000000
--- a/libctru/source/linear.c
+++ /dev/null
@@ -1,32 +0,0 @@
-#include <3ds.h>
-
-extern u32 __linear_heap, __linear_heap_size;
-
-// TODO: this allocator sucks! It is not thread-safe and you cannot 'free' this memory.
-void* linearAlloc(size_t size)
-{
-	static size_t currentOffset = 0;
-	size_t free = __linear_heap_size - currentOffset;
-
-	// Enforce 16-byte alignment
-	size = (size + 15) &~ 15;
-
-	void* mem = NULL;
-	if (free >= size)
-	{
-		mem = (void*)(__linear_heap + currentOffset);
-		currentOffset += size;
-	}
-
-	return mem;
-}
-
-void* linearRealloc(void* mem, size_t size)
-{
-	return NULL; // TODO
-}
-
-void linearFree(void* mem)
-{
-	// TODO
-}

From 1f413a7d4485b8e1d1a69d32c83aa703f2c17898 Mon Sep 17 00:00:00 2001
From: StapleButter <thetotalworm@gmail.com>
Date: Thu, 18 Sep 2014 22:09:15 +0200
Subject: [PATCH 13/14] Add synchronization mechanism for entering sleep mode.

When the APT status is APP_PREPARE_SLEEPMODE, the application main thread should call aptSignalReadyForSleep() to signal that it is ready to enter sleep mode, and then call aptWaitStatusEvent() as usual.

Example code:

APP_STATUS status;
while ((status = aptGetStatus()) != APP_EXITING)
{
	if(status==APP_RUNNING)
	{
		// application logic here
	}
	else if(status == APP_SUSPENDING)
	{
		aptReturnToMenu();
	}
	else if(status == APP_PREPARE_SLEEPMODE)
	{
		aptSignalReadyForSleep();
		aptWaitStatusEvent();
	}
}

This maybe isn't the proper/recommended way to do sleep mode, but I tested it multiple times and it always worked reliably.

(note: maybe the sample code above will not work if GPU drawing is done in a separate thread, haven't tested that)
---
 libctru/include/3ds/services/apt.h |  1 +
 libctru/source/services/apt.c      | 20 ++++++++++++++++----
 2 files changed, 17 insertions(+), 4 deletions(-)

diff --git a/libctru/include/3ds/services/apt.h b/libctru/include/3ds/services/apt.h
index 9377c0b..b8c05d9 100644
--- a/libctru/include/3ds/services/apt.h
+++ b/libctru/include/3ds/services/apt.h
@@ -48,6 +48,7 @@ u32 aptGetStatusPower();//This can be used when the status is APP_SUSPEND* to ch
 void aptSetStatusPower(u32 status);
 void aptReturnToMenu();//This should be called by the user application when aptGetStatus() returns APP_SUSPENDING, not calling this will result in return-to-menu being disabled with the status left at APP_SUSPENDING. This function will not return until the system returns to the application, or when the status was changed to APP_EXITING.
 void aptWaitStatusEvent();
+void aptSignalReadyForSleep();
 NS_APPID aptGetMenuAppID();
 
 Result APT_GetLockHandle(Handle* handle, u16 flags, Handle* lockHandle);
diff --git a/libctru/source/services/apt.c b/libctru/source/services/apt.c
index d2f65d9..3c8de84 100644
--- a/libctru/source/services/apt.c
+++ b/libctru/source/services/apt.c
@@ -26,6 +26,7 @@ Handle aptStatusEvent = 0;
 APP_STATUS aptStatus = APP_NOTINITIALIZED;
 APP_STATUS aptStatusBeforeSleep = APP_NOTINITIALIZED;
 u32 aptStatusPower = 0;
+Handle aptSleepSync = 0;
 
 u32 aptParameters[0x1000/4]; //TEMP
 
@@ -204,21 +205,24 @@ static void __handle_notification() {
 	case APTSIGNAL_PREPARESLEEP:
 		// Reply to sleep-request.
 		aptStatusBeforeSleep = aptGetStatus();
+		aptSetStatus(APP_PREPARE_SLEEPMODE);
+		svcWaitSynchronization(aptSleepSync, U64_MAX);
+		svcClearEvent(aptSleepSync);
+		
 		aptOpenSession();
 		APT_ReplySleepQuery(NULL, currentAppId, 0x1);
 		aptCloseSession();
-
-		aptSetStatus(APP_PREPARE_SLEEPMODE);
 		break;
 
 	case APTSIGNAL_ENTERSLEEP:
 		if(aptGetStatus() == APP_PREPARE_SLEEPMODE)
 		{
 			// Report into sleep-mode.
+			aptSetStatus(APP_SLEEPMODE);
+			
 			aptOpenSession();
 			APT_ReplySleepNotificationComplete(NULL, currentAppId);
 			aptCloseSession();
-			aptSetStatus(APP_SLEEPMODE);
 		}
 		break;
 
@@ -319,6 +323,7 @@ Result aptInit(void)
 	}
 
 	svcCreateEvent(&aptStatusEvent, 0);
+	svcCreateEvent(&aptSleepSync, 0);
 	return 0;
 }
 
@@ -344,6 +349,8 @@ void aptExit()
 		APT_CloseApplication(NULL, 0x0, 0x0, 0x0);
 		aptCloseSession();
 	}
+	
+	svcCloseHandle(aptSleepSync);
 
 	svcCloseHandle(aptStatusMutex);
 	//svcCloseHandle(aptLockHandle);
@@ -421,7 +428,7 @@ void aptSetStatus(APP_STATUS status)
 
 	svcWaitSynchronization(aptStatusMutex, U64_MAX);
 
-	prevstatus = status;
+	prevstatus = aptStatus;
 	aptStatus = status;
 
 	if(prevstatus != APP_NOTINITIALIZED)
@@ -461,6 +468,11 @@ void aptCloseSession()
 	svcReleaseMutex(aptLockHandle);
 }
 
+void aptSignalReadyForSleep()
+{
+	svcSignalEvent(aptSleepSync);
+}
+
 Result APT_GetLockHandle(Handle* handle, u16 flags, Handle* lockHandle)
 {
 	if(!handle)handle=&aptuHandle;

From af31a7c861e472ff9ef9e0def347a66aeb51a61e Mon Sep 17 00:00:00 2001
From: StapleButter <thetotalworm@gmail.com>
Date: Thu, 23 Oct 2014 17:56:56 +0200
Subject: [PATCH 14/14] * add code for alphablending, color logic op, alpha
 test and multitexturing. * add GPU_FinishDrawing() to be called after a batch
 of GPU_DrawArray() calls if you're done drawing or if you intend to change
 the GPU configuration before drawing more. Also fix GPU_Finalize(). With
 those changes, the GPU no longer freezes if you call GPU_DrawArray() an even
 number of times. * fix GPU_SetViewport() to allow color buffer reading, so
 blending and logicop work as expected.

---
 libctru/include/3ds/gpu/gpu.h | 71 +++++++++++++++++++++++++++++++++-
 libctru/source/gpu/gpu.c      | 72 +++++++++++++++++++++++++++++++----
 2 files changed, 135 insertions(+), 8 deletions(-)

diff --git a/libctru/include/3ds/gpu/gpu.h b/libctru/include/3ds/gpu/gpu.h
index 0b2ee2a..aaceb0a 100644
--- a/libctru/include/3ds/gpu/gpu.h
+++ b/libctru/include/3ds/gpu/gpu.h
@@ -9,6 +9,13 @@ void GPUCMD_Add(u32 cmd, u32* param, u32 paramlength);
 void GPUCMD_AddSingleParam(u32 cmd, u32 param);
 void GPUCMD_Finalize();
 
+typedef enum
+{
+	GPU_TEXUNIT0 = 0x1,
+	GPU_TEXUNIT1 = 0x2,
+	GPU_TEXUNIT2 = 0x4
+} GPU_TEXUNIT;
+
 typedef enum{
 	GPU_RGBA8=0x0,
 	GPU_RGB8=0x1,
@@ -37,6 +44,54 @@ typedef enum
 	GPU_GEQUAL = 7
 }GPU_TESTFUNC;
 
+typedef enum
+{
+	GPU_BLEND_ADD = 0,
+	GPU_BLEND_SUBTRACT = 1,
+	GPU_BLEND_REVERSE_SUBTRACT = 2,
+	GPU_BLEND_MIN = 3,
+	GPU_BLEND_MAX = 4
+} GPU_BLENDEQUATION;
+
+typedef enum
+{
+	GPU_ZERO = 0,
+	GPU_ONE = 1,
+	GPU_SRC_COLOR = 2,
+	GPU_ONE_MINUS_SRC_COLOR = 3,
+	GPU_DST_COLOR = 4,
+	GPU_ONE_MINUS_DST_COLOR = 5,
+	GPU_SRC_ALPHA = 6,
+	GPU_ONE_MINUS_SRC_ALPHA = 7,
+	GPU_DST_ALPHA = 8,
+	GPU_ONE_MINUS_DST_ALPHA = 9,
+	GPU_CONSTANT_COLOR = 10,
+	GPU_ONE_MINUS_CONSTANT_COLOR = 11,
+	GPU_CONSTANT_ALPHA = 12,
+	GPU_ONE_MINUS_CONSTANT_ALPHA = 13,
+	GPU_SRC_ALPHA_SATURATE = 14
+} GPU_BLENDFACTOR;
+
+typedef enum
+{
+	GPU_LOGICOP_CLEAR = 0,
+	GPU_LOGICOP_AND = 1,
+	GPU_LOGICOP_AND_REVERSE = 2,
+	GPU_LOGICOP_COPY = 3,
+	GPU_LOGICOP_SET = 4,
+	GPU_LOGICOP_COPY_INVERTED = 5,
+	GPU_LOGICOP_NOOP = 6,
+	GPU_LOGICOP_INVERT = 7,
+	GPU_LOGICOP_NAND = 8,
+	GPU_LOGICOP_OR = 9,
+	GPU_LOGICOP_NOR = 10,
+	GPU_LOGICOP_XOR = 11,
+	GPU_LOGICOP_EQUIV = 12,
+	GPU_LOGICOP_AND_INVERTED = 13,
+	GPU_LOGICOP_OR_REVERSE = 14,
+	GPU_LOGICOP_OR_INVERTED = 15
+} GPU_LOGICOP;
+
 typedef enum{
 	GPU_BYTE = 0,
 	GPU_UNSIGNED_BYTE = 1,
@@ -84,14 +139,28 @@ typedef enum{
 }GPU_Primitive_t;
 
 void GPU_SetUniform(u32 startreg, u32* data, u32 numreg);
+
 void GPU_SetViewport(u32* depthBuffer, u32* colorBuffer, u32 x, u32 y, u32 w, u32 h);
+
 void GPU_DepthRange(float nearVal, float farVal);
+void GPU_SetAlphaTest(bool enable, GPU_TESTFUNC function, u8 ref);
 void GPU_SetDepthTest(bool enable, GPU_TESTFUNC function, u8 ref);
 void GPU_SetStencilTest(bool enable, GPU_TESTFUNC function, u8 ref);
 void GPU_SetFaceCulling(GPU_CULLMODE mode);
+
+// these two can't be used together
+void GPU_SetAlphaBlending(GPU_BLENDEQUATION colorEquation, GPU_BLENDEQUATION alphaEquation, 
+	GPU_BLENDFACTOR colorSrc, GPU_BLENDFACTOR colorDst, 
+	GPU_BLENDFACTOR alphaSrc, GPU_BLENDFACTOR alphaDst);
+void GPU_SetColorLogicOp(GPU_LOGICOP op);
+
 void GPU_SetAttributeBuffers(u8 totalAttributes, u32* baseAddress, u64 attributeFormats, u16 attributeMask, u64 attributePermutation, u8 numBuffers, u32 bufferOffsets[], u64 bufferPermutations[], u8 bufferNumAttributes[]);
-void GPU_SetTexture(u32* data, u16 width, u16 height, u32 param, GPU_TEXCOLOR colorType);
+
+void GPU_SetTextureEnable(GPU_TEXUNIT units); // GPU_TEXUNITx values can be ORed together to enable multiple texture units
+void GPU_SetTexture(GPU_TEXUNIT unit, u32* data, u16 width, u16 height, u32 param, GPU_TEXCOLOR colorType);
 void GPU_SetTexEnv(u8 id, u16 rgbSources, u16 alphaSources, u16 rgbOperands, u16 alphaOperands, GPU_COMBINEFUNC rgbCombine, GPU_COMBINEFUNC alphaCombine, u32 constantColor);
 
 void GPU_DrawArray(GPU_Primitive_t primitive, u32 n);
 void GPU_DrawElements(GPU_Primitive_t primitive, u32* indexArray, u32 n);
+
+void GPU_FinishDrawing();
diff --git a/libctru/source/gpu/gpu.c b/libctru/source/gpu/gpu.c
index d6685e3..835f1ff 100644
--- a/libctru/source/gpu/gpu.c
+++ b/libctru/source/gpu/gpu.c
@@ -65,6 +65,9 @@ void GPUCMD_AddSingleParam(u32 cmd, u32 param)
 
 void GPUCMD_Finalize()
 {
+	GPUCMD_AddSingleParam(0x0008025E, 0x00000000);
+	GPUCMD_AddSingleParam(0x000F0111, 0x00000001);
+	GPUCMD_AddSingleParam(0x000F0110, 0x00000001);
 	GPUCMD_AddSingleParam(0x000F0010, 0x12345678);
 }
 
@@ -272,7 +275,7 @@ void GPU_SetViewport(u32* depthBuffer, u32* colorBuffer, u32 x, u32 y, u32 w, u3
 	GPUCMD_Add(0x800F0065, param, 0x00000003);
 
 	//enable depth buffer
-	param[0x0]=0x00000000;
+	param[0x0]=0x0000000F;
 	param[0x1]=0x0000000F;
 	param[0x2]=0x00000002;
 	param[0x3]=0x00000002;
@@ -286,6 +289,11 @@ void GPU_DepthRange(float nearVal, float farVal)
 	GPUCMD_AddSingleParam(0x000F004E, f32tof24(farVal));
 }
 
+void GPU_SetAlphaTest(bool enable, GPU_TESTFUNC function, u8 ref)
+{
+	GPUCMD_AddSingleParam(0x000F0104, (enable&1)|((function&7)<<4)|(ref<<8));
+}
+
 void GPU_SetStencilTest(bool enable, GPU_TESTFUNC function, u8 ref)
 {
 	GPUCMD_AddSingleParam(0x000F0105, (enable&1)|((function&7)<<4)|(ref<<8));
@@ -296,12 +304,54 @@ void GPU_SetDepthTest(bool enable, GPU_TESTFUNC function, u8 ref)
 	GPUCMD_AddSingleParam(0x000F0107, (enable&1)|((function&7)<<4)|(ref<<8));
 }
 
-void GPU_SetTexture(u32* data, u16 width, u16 height, u32 param, GPU_TEXCOLOR colorType)
+void GPU_SetAlphaBlending(GPU_BLENDEQUATION colorEquation, GPU_BLENDEQUATION alphaEquation, 
+	GPU_BLENDFACTOR colorSrc, GPU_BLENDFACTOR colorDst, 
+	GPU_BLENDFACTOR alphaSrc, GPU_BLENDFACTOR alphaDst)
 {
-	GPUCMD_AddSingleParam(0x000F008E, colorType);
-	GPUCMD_AddSingleParam(0x000F0085, ((u32)data)>>3);
-	GPUCMD_AddSingleParam(0x000F0082, (width)|(height<<16));
-	GPUCMD_AddSingleParam(0x000F0083, param);
+	// TODO: fixed color
+	// it is controlled by command 0103 but I haven't found were to place said command without freezing the GPU
+	
+	GPUCMD_AddSingleParam(0x000F0101, colorEquation | (alphaEquation<<8) | (colorSrc<<16) | (colorDst<<20) | (alphaSrc<<24) | (alphaDst<<28));
+	GPUCMD_AddSingleParam(0x00020100, 0x00000100);
+}
+
+void GPU_SetColorLogicOp(GPU_LOGICOP op)
+{
+	GPUCMD_AddSingleParam(0x000F0102, op);
+	GPUCMD_AddSingleParam(0x00020100, 0x00000000);
+}
+
+void GPU_SetTextureEnable(GPU_TEXUNIT units)
+{
+	GPUCMD_AddSingleParam(0x0002006F, units<<8); 			// enables texcoord outputs
+	GPUCMD_AddSingleParam(0x000F0080, 0x00011000|units);	// enables texture units
+}
+
+void GPU_SetTexture(GPU_TEXUNIT unit, u32* data, u16 width, u16 height, u32 param, GPU_TEXCOLOR colorType)
+{
+	switch (unit)
+	{
+	case GPU_TEXUNIT0:
+		GPUCMD_AddSingleParam(0x000F008E, colorType);
+		GPUCMD_AddSingleParam(0x000F0085, ((u32)data)>>3);
+		GPUCMD_AddSingleParam(0x000F0082, (width)|(height<<16));
+		GPUCMD_AddSingleParam(0x000F0083, param);
+		break;
+		
+	case GPU_TEXUNIT1:
+		GPUCMD_AddSingleParam(0x000F0096, colorType);
+		GPUCMD_AddSingleParam(0x000F0095, ((u32)data)>>3);
+		GPUCMD_AddSingleParam(0x000F0092, (width)|(height<<16));
+		GPUCMD_AddSingleParam(0x000F0093, param);
+		break;
+		
+	case GPU_TEXUNIT2:
+		GPUCMD_AddSingleParam(0x000F009E, colorType);
+		GPUCMD_AddSingleParam(0x000F009D, ((u32)data)>>3);
+		GPUCMD_AddSingleParam(0x000F009A, (width)|(height<<16));
+		GPUCMD_AddSingleParam(0x000F009B, param);
+		break;
+	}
 }
 
 const u8 GPU_FORMATSIZE[4]={1,1,2,4};
@@ -385,7 +435,6 @@ void GPU_DrawArray(GPU_Primitive_t primitive, u32 n)
 	GPUCMD_AddSingleParam(0x000F0231, 0x00000001);
 
 	GPUCMD_AddSingleParam(0x000F0111, 0x00000001);
-	GPUCMD_AddSingleParam(0x000F0110, 0x00000001);
 }
 
 void GPU_DrawElements(GPU_Primitive_t primitive, u32* indexArray, u32 n)
@@ -405,4 +454,13 @@ void GPU_DrawElements(GPU_Primitive_t primitive, u32* indexArray, u32 n)
 	GPUCMD_AddSingleParam(0x000F022F, 0x00000001);
 	GPUCMD_AddSingleParam(0x00010245, 0x00000001);
 	GPUCMD_AddSingleParam(0x000F0231, 0x00000001);
+	
+	// CHECKME: does this one also require command 0x0111 at the end?
+}
+
+void GPU_FinishDrawing()
+{
+	GPUCMD_AddSingleParam(0x000F0111, 0x00000001);
+	GPUCMD_AddSingleParam(0x000F0110, 0x00000001); 
+	GPUCMD_AddSingleParam(0x000F0063, 0x00000001);
 }