added the possibility to allocate aligned memory, an use it to allocate the buffer utilized in texture decoding, this will make a little easy to use aligned writes when possible in sse2/3 optimized algorithms.

some code additions for future use ;).
please gcc user test this as i don't have opportunity to test it myself i only use reference code to.

git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@7247 8ced0084-cf51-0410-be5f-012b33b47a6e

Source/Core/Common/Src/ABI.cpp

@@ -124,6 +124,15 @@ void XEmitter::ABI_CallFunctionCCCP(void *func, u32 param1, u32 param2,u32 param
 	ABI_RestoreStack(4 * 4);
 }
 
+void XEmitter::ABI_CallFunctionPPC(void *func, void *param1, void *param2,u32 param3) {
+	ABI_AlignStack(3 * 4);
+	PUSH(32, Imm32(param3));
+	PUSH(32, Imm32((u32)param2));
+	PUSH(32, Imm32((u32)param1));
+	CALL(func);
+	ABI_RestoreStack(3 * 4);
+}
+
 // Pass a register as a parameter.
 void XEmitter::ABI_CallFunctionR(void *func, X64Reg reg1) {
 	ABI_AlignStack(1 * 4);
@@ -326,6 +335,21 @@ void XEmitter::ABI_CallFunctionCCCP(void *func, u32 param1, u32 param2, u32 para
 	}
 }
 
+void XEmitter::ABI_CallFunctionPPC(void *func, void *param1, void *param2, u32 param3) {
+	MOV(64, R(ABI_PARAM1), Imm64((u64)param1));
+	MOV(64, R(ABI_PARAM2), Imm64((u64)param2));
+	MOV(32, R(ABI_PARAM3), Imm32(param3));
+	u64 distance = u64(func) - (u64(code) + 5);
+	if (distance >= 0x0000000080000000ULL
+	 && distance <  0xFFFFFFFF80000000ULL) {
+	    // Far call
+	    MOV(64, R(RAX), Imm64((u64)func));
+	    CALLptr(R(RAX));
+	} else {
+	    CALL(func);
+	}
+}
+
 // Pass a register as a parameter.
 void XEmitter::ABI_CallFunctionR(void *func, X64Reg reg1) {
 	if (reg1 != ABI_PARAM1)

Source/Core/Common/Src/MemoryUtil.cpp

@@ -74,19 +74,54 @@ void* AllocateMemoryPages(size_t size)
 	return ptr;
 }
 
-void FreeMemoryPages(void* ptr, size_t size)
+void* AllocateAlignedMemory(size_t size,size_t alignment)
 {
 #ifdef _WIN32
+	void* ptr =  _aligned_malloc(size,alignment);
+#else
+	void* ptr = NULL;
+	posix_memalign(&ptr, alignment, size);
+;
+#endif
+
+	// printf("Mapped memory at %p (size %ld)\n", ptr,
+	//	(unsigned long)size);
+
+	if (ptr == NULL)
+		PanicAlert("Failed to allocate aligned memory");
+
+	return ptr;
+}
+
+void FreeMemoryPages(void* ptr, size_t size)
+{
 	if (ptr)
 	{
+#ifdef _WIN32
+	
 		if (!VirtualFree(ptr, 0, MEM_RELEASE))
 			PanicAlert("FreeMemoryPages failed!\n%s", GetLastErrorMsg());
 		ptr = NULL; // Is this our responsibility?
-	}
+	
 #else
 		munmap(ptr, size);
 #endif
 	}
+}
+
+void FreeAlignedMemory(void* ptr)
+{
+	if (ptr)
+	{
+#ifdef _WIN32
+	
+		_aligned_free(ptr);
+	
+#else
+	free(ptr);
+#endif
+	}
+}
 
 void WriteProtectMemory(void* ptr, size_t size, bool allowExecute)
 {

Source/Core/Common/Src/MemoryUtil.h

@@ -26,6 +26,8 @@
 void* AllocateExecutableMemory(size_t size, bool low = true);
 void* AllocateMemoryPages(size_t size);
 void FreeMemoryPages(void* ptr, size_t size);
+void* AllocateAlignedMemory(size_t size,size_t alignment);
+void FreeAlignedMemory(void* ptr);
 void WriteProtectMemory(void* ptr, size_t size, bool executable = false);
 void UnWriteProtectMemory(void* ptr, size_t size, bool allowExecute = false);
 std::string MemUsage();

Source/Core/Common/Src/x64Emitter.h

@@ -625,6 +625,7 @@ public:
 	void ABI_CallFunctionCCC(void *func, u32 param1, u32 param2, u32 param3);
 	void ABI_CallFunctionCCP(void *func, u32 param1, u32 param2, void *param3);
 	void ABI_CallFunctionCCCP(void *func, u32 param1, u32 param2,u32 param3, void *param4);
+	void ABI_CallFunctionPPC(void *func, void *param1, void *param2,u32 param3);
 	void ABI_CallFunctionAC(void *func, const Gen::OpArg &arg1, u32 param2);
 	void ABI_CallFunctionA(void *func, const Gen::OpArg &arg1);
 

Source/Core/VideoCommon/Src/Statistics.h

@@ -90,6 +90,7 @@ extern Statistics stats;
 
 #ifdef STATISTICS
 #define INCSTAT(a) (a)++;
+#define DECSTAT(a) (a)--;
 #define ADDSTAT(a,b) (a)+=(b);
 #define SETSTAT(a,x) (a)=(int)(x);
 #define SETSTAT_UINT(a,x) (a)=(u32)(x);

Source/Core/VideoCommon/Src/TextureCacheBase.cpp

@@ -23,7 +23,7 @@ enum
 
 TextureCache *g_texture_cache;
 
-u8 *TextureCache::temp = NULL;
+ GC_ALIGNED16(u8 *TextureCache::temp) = NULL;
 
 TextureCache::TexCache TextureCache::textures;
 bool TextureCache::DeferredInvalidate;
@@ -44,7 +44,7 @@ TextureCache::TCacheEntryBase::~TCacheEntryBase()
 TextureCache::TextureCache()
 {
 	if (!temp)
-		temp = (u8*)AllocateMemoryPages(TEMP_SIZE);
+		temp =(u8*) AllocateAlignedMemory(TEMP_SIZE,16);
 	TexDecoder_SetTexFmtOverlayOptions(g_ActiveConfig.bTexFmtOverlayEnable, g_ActiveConfig.bTexFmtOverlayCenter);
     if(g_ActiveConfig.bHiresTextures && !g_ActiveConfig.bDumpTextures)
 		HiresTextures::Init(SConfig::GetInstance().m_LocalCoreStartupParameter.m_strUniqueID.c_str());
@@ -81,7 +81,7 @@ TextureCache::~TextureCache()
 	Invalidate(true);
 	if (temp)
 	{
-		FreeMemoryPages(temp, TEMP_SIZE);
+		FreeAlignedMemory(temp);
 		temp = NULL;
 	}
 }

Source/Core/VideoCommon/Src/TextureCacheBase.h

@@ -95,7 +95,7 @@ public:
 protected:
 	TextureCache();
 
-	static u8 *temp;
+	static  GC_ALIGNED16(u8 *temp);
 
 private:
 	typedef std::map<u32, TCacheEntryBase*> TexCache;

Source/Core/VideoCommon/Src/TextureDecoder.cpp

@@ -44,7 +44,7 @@ extern const unsigned char sfont_raw[][9*10];
 
 // TRAM
 // STATE_TO_SAVE
-u8 texMem[TMEM_SIZE];
+ GC_ALIGNED16(u8 texMem[TMEM_SIZE]);
 
 
 // Gamecube/Wii texture decoder

Source/Core/VideoCommon/Src/TextureDecoder.h

@@ -23,7 +23,7 @@ enum
     TMEM_SIZE = 1024*1024,
     HALFTMEM_SIZE = 512*1024
 };
-extern u8 texMem[TMEM_SIZE];
+extern  GC_ALIGNED16(u8 texMem[TMEM_SIZE]);
 
 enum TextureFormat
 {