Merge pull request #64 from Nekotekina/master

Last changes
2024-11-22 10:42:36 +01:00 · 2014-02-09 14:50:35 +01:00 · 2014-02-09 14:50:35 +01:00 · 4ebd974f5e
commit 4ebd974f5e
parent 66a9acfb04 3308fefa7f
51 changed files with 1218 additions and 618 deletions
--- a/Utilities/BEType.h
+++ b/Utilities/BEType.h
@ -2,6 +2,10 @@

 #include "Utilities/GNU.h"

+#define se16(x) const_se_t<u16, x>::value
+#define se32(x) const_se_t<u32, x>::value
+#define se64(x) const_se_t<u64, x>::value
+
 template<typename T, int size = sizeof(T)> struct se_t;
 template<typename T> struct se_t<T, 1> { static __forceinline void func(T& dst, const T src) { (u8&)dst = (u8&)src; } };
 template<typename T> struct se_t<T, 2> { static __forceinline void func(T& dst, const T src) { (u16&)dst = _byteswap_ushort((u16&)src); } };
--- a/Utilities/SMutex.cpp
+++ b/Utilities/SMutex.cpp
@ -0,0 +1,26 @@
+#include <stdafx.h>
+#include <Utilities/SMutex.h>
+
+__forceinline void SM_Sleep()
+{
+	Sleep(1);
+}
+
+__forceinline DWORD SM_GetCurrentThreadId()
+{
+	return GetCurrentThreadId();
+}
+
+__forceinline u32 SM_GetCurrentCPUThreadId()
+{
+	if (CPUThread* t = GetCurrentCPUThread())
+	{
+		return t->GetId();
+	}
+	return 0;
+}
+
+__forceinline be_t<u32> SM_GetCurrentCPUThreadIdBE()
+{
+	return SM_GetCurrentCPUThreadId();
+}
--- a/Utilities/SMutex.h
+++ b/Utilities/SMutex.h
@ -0,0 +1,155 @@
+#pragma once
+#include <atomic>
+
+extern void SM_Sleep();
+extern DWORD SM_GetCurrentThreadId();
+extern u32 SM_GetCurrentCPUThreadId();
+extern be_t<u32> SM_GetCurrentCPUThreadIdBE();
+
+enum SMutexResult
+{
+	SMR_OK = 0, // succeeded (lock, trylock, unlock)
+	SMR_FAILED, // failed (trylock, unlock)
+	SMR_DEADLOCK, // mutex reached deadlock (lock, trylock)
+	SMR_SIGNAL = SMR_DEADLOCK, // unlock can be used for signaling specific thread
+	SMR_PERMITTED, // not owner of the mutex (unlock)
+	SMR_ABORT, // emulator has been stopped (lock, trylock, unlock)
+	SMR_DESTROYED, // mutex has been destroyed (lock, trylock, unlock)
+	SMR_TIMEOUT, // timed out (lock)
+};
+
+template
+<
+	typename T,
+	u32 free_value = 0,
+	u32 dead_value = ~0,
+	void (wait)() = SM_Sleep
+>
+class SMutexBase
+{
+	static_assert(sizeof(T) == 4, "Invalid SMutexBase typename");
+	std::atomic<T> owner;
+
+public:
+	SMutexBase()
+		: owner((T)free_value)
+	{
+	}
+
+	~SMutexBase()
+	{
+		lock((T)dead_value);
+		owner = (T)dead_value;
+	}
+
+	__forceinline T GetOwner() const
+	{
+		return (T&)owner;
+	}
+
+	SMutexResult trylock(T tid)
+	{
+		T old = (T)free_value;
+
+		if (!owner.compare_exchange_strong(old, tid))
+		{
+			if (old == tid)
+			{
+				return SMR_DEADLOCK;
+			}
+			if (Emu.IsStopped())
+			{
+				return SMR_ABORT;
+			}
+			if (old == (T)dead_value)
+			{
+				return SMR_DESTROYED;
+			}
+
+			return SMR_FAILED;
+		}
+
+		return SMR_OK;
+	}
+
+	SMutexResult unlock(T tid, T to = (T)free_value)
+	{
+		T old = tid;
+
+		if (!owner.compare_exchange_strong(old, to))
+		{
+			if (old == (T)free_value)
+			{
+				return SMR_FAILED;
+			}
+			if (old == (T)dead_value)
+			{
+				return SMR_DESTROYED;
+			}
+
+			return SMR_PERMITTED;
+		}
+
+		return SMR_OK;
+	}
+
+	SMutexResult lock(T tid, u64 timeout = 0)
+	{
+		u64 counter = 0;
+
+		while (true)
+		{
+			switch (SMutexResult res = trylock(tid))
+			{
+				case SMR_FAILED: break;
+				default: return res;
+			}
+
+			wait();
+
+			if (timeout && counter++ > timeout)
+			{
+				return SMR_TIMEOUT;
+			}
+		}
+	}
+};
+
+template<typename T, T (get_tid)()>
+class SMutexLockerBase
+{
+	SMutexBase<T>& sm;
+public:
+	const T tid;
+
+	SMutexLockerBase(SMutexBase<T>& _sm)
+		: sm(_sm)
+		, tid(get_tid())
+	{
+		if (!tid)
+		{
+			ConLog.Error("SMutexLockerBase: thread id == 0");
+			Emu.Pause();
+		}
+		sm.lock(tid);
+	}
+
+	~SMutexLockerBase()
+	{
+		sm.unlock(tid);
+	}
+};
+
+typedef SMutexBase<DWORD>
+	SMutexGeneral;
+typedef SMutexBase<u32>
+	SMutex;
+typedef SMutexBase<be_t<u32>>
+	SMutexBE;
+
+typedef SMutexLockerBase<DWORD, SM_GetCurrentThreadId>
+	SMutexGeneralLocker;
+typedef SMutexLockerBase<u32, SM_GetCurrentCPUThreadId>
+	SMutexLocker;
+typedef SMutexLockerBase<be_t<u32>, SM_GetCurrentCPUThreadIdBE>
+	SMutexBELocker;
--- a/rpcs3/Emu/CPU/CPUThread.cpp
+++ b/rpcs3/Emu/CPU/CPUThread.cpp
@ -1,6 +1,8 @@
 #include "stdafx.h"
 #include "CPUThread.h"

+reservation_struct reservation;
+
 CPUThread* GetCurrentCPUThread()
 {
 	return (CPUThread*)GetCurrentNamedThread();
@ -89,14 +91,12 @@ void CPUThread::Wait(const CPUThread& thr)

 bool CPUThread::Sync()
 {
-	wxCriticalSectionLocker lock(m_cs_sync);
-
 	return m_sync_wait;
 }

 int CPUThread::ThreadStatus()
 {
-	if(Emu.IsStopped())
+	if(Emu.IsStopped() || IsStopped() || IsPaused())
 	{
 		return CPUThread_Stopped;
 	}
@ -236,7 +236,7 @@ void CPUThread::Pause()
 	DoPause();
 	Emu.CheckStatus();

-	ThreadBase::Stop();
+	// ThreadBase::Stop(); // "Abort() called" exception
 #ifndef QT_UI
 	wxGetApp().SendDbgCommand(DID_PAUSED_THREAD, this);
 #endif
--- a/rpcs3/Emu/CPU/CPUThread.h
+++ b/rpcs3/Emu/CPU/CPUThread.h
@ -1,6 +1,25 @@
 #pragma once
 #include "Emu/Memory/MemoryBlock.h"
 #include "Emu/CPU/CPUDecoder.h"
+#include "Utilities/SMutex.h"
+
+struct reservation_struct
+{
+	SMutex mutex; // mutex for updating reservation_owner and data
+	volatile u32 owner; // id of thread that got reservation
+	volatile u32 addr;
+	volatile u32 size;
+	volatile u32 data32;
+	volatile u64 data64;
+	// atm, PPU can't break SPU MFC reservation correctly
+
+	__forceinline void clear()
+	{
+		owner = 0;
+	}
+};
+
+extern reservation_struct reservation;

 enum CPUThreadType
 {
--- a/rpcs3/Emu/CPU/CPUThreadManager.cpp
+++ b/rpcs3/Emu/CPU/CPUThreadManager.cpp
@ -76,6 +76,8 @@ void CPUThreadManager::RemoveThread(const u32 id)

 s32 CPUThreadManager::GetThreadNumById(CPUThreadType type, u32 id)
 {
+	std::lock_guard<std::mutex> lock(m_mtx_thread);
+
 	s32 num = 0;

 	for(u32 i=0; i<m_threads.GetCount(); ++i)
@ -89,16 +91,17 @@ s32 CPUThreadManager::GetThreadNumById(CPUThreadType type, u32 id)

 CPUThread* CPUThreadManager::GetThread(u32 id)
 {
-	for(u32 i=0; i<m_threads.GetCount(); ++i)
-	{
-		if(m_threads[i].GetId() == id) return &m_threads[i];
-	}
+	CPUThread* res;

-	return nullptr;
+	if (!Emu.GetIdManager().GetIDData(id, res)) return nullptr;
+
+	return res;
 }

 void CPUThreadManager::Exec()
 {
+	std::lock_guard<std::mutex> lock(m_mtx_thread);
+
 	for(u32 i=0; i<m_threads.GetCount(); ++i)
 	{
 		m_threads[i].Exec();
--- a/rpcs3/Emu/Cell/PPCThread.cpp
+++ b/rpcs3/Emu/Cell/PPCThread.cpp
@ -16,6 +16,7 @@ PPCThread* GetCurrentPPCThread()

 PPCThread::PPCThread(CPUThreadType type) : CPUThread(type)
 {
+	memset(m_args, 0, sizeof(m_args));
 }

 PPCThread::~PPCThread()
@ -24,14 +25,13 @@ PPCThread::~PPCThread()

 void PPCThread::DoReset()
 {
-	memset(m_args, 0, sizeof(u64) * 4);
 }

 void PPCThread::InitStack()
 {
 	if(m_stack_addr) return;
 	if(m_stack_size == 0) m_stack_size = 0x10000;
-	m_stack_addr = Memory.StackMem.Alloc(Memory.AlignAddr(m_stack_size, 0x100));
+	m_stack_addr = Memory.StackMem.AllocAlign(m_stack_size, 0x100);

 	m_stack_point = m_stack_addr + m_stack_size;
 	/*
--- a/rpcs3/Emu/Cell/PPUInterpreter.h
+++ b/rpcs3/Emu/Cell/PPUInterpreter.h
@ -1656,7 +1656,7 @@ private:
 		{
 			CPU.VPR[vd]._u8[15] = CPU.VPR[va]._u8[15] >> sh;

-			for (uint b = 14; b >= 0; b--)
+			for (uint b = 14; ~b; b--)
 			{
 				CPU.VPR[vd]._u8[b] = (CPU.VPR[va]._u8[b] >> sh) | (CPU.VPR[va]._u8[b+1] << (8 - sh));
 			}
@ -2362,9 +2362,12 @@ private:
 	void LWARX(u32 rd, u32 ra, u32 rb)
 	{
 		const u64 addr = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];
-		CPU.reserve_addr = addr;
-		CPU.reserve = true;
-		CPU.GPR[rd] = Memory.Read32(addr);
+
+		SMutexLocker lock(reservation.mutex);
+		reservation.owner = lock.tid;
+		reservation.addr = addr;
+		reservation.size = 4;
+		reservation.data32 = CPU.GPR[rd] = Memory.Read32(addr);
 	}
 	void LDX(u32 rd, u32 ra, u32 rb)
 	{
@ -2535,9 +2538,12 @@ private:
 	void LDARX(u32 rd, u32 ra, u32 rb)
 	{
 		const u64 addr = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];
-		CPU.reserve_addr = addr;
-		CPU.reserve = true;
-		CPU.GPR[rd] = Memory.Read64(addr);
+
+		SMutexLocker lock(reservation.mutex);
+		reservation.owner = lock.tid;
+		reservation.addr = addr;
+		reservation.size = 8;
+		reservation.data64 = CPU.GPR[rd] = Memory.Read64(addr);
 	}
 	void DCBF(u32 ra, u32 rb)
 	{
@ -2650,25 +2656,19 @@ private:
 	}
 	void STWCX_(u32 rs, u32 ra, u32 rb)
 	{
-		CPU.SetCR(0, CPU.XER.SO ? CR_SO : 0);
-		
-		if(CPU.reserve)
-		{
-			const u64 addr = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];
+		const u64 addr = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];

-			if(addr == CPU.reserve_addr)
-			{
-				Memory.Write32(addr, CPU.GPR[rs]);
-				CPU.SetCR_EQ(0, true);
-				CPU.reserve = false;
-			}
-			else
-			{
-				static const bool u = 0;
-				if(u) Memory.Write32(addr, CPU.GPR[rs]);
-				CPU.SetCR_EQ(0, u);
-				CPU.reserve = false;
-			}
+		SMutexLocker lock(reservation.mutex);
+		if (lock.tid == reservation.owner && reservation.addr == addr && reservation.size == 4)
+		{
+			// Memory.Write32(addr, CPU.GPR[rs]);
+			CPU.SetCR_EQ(0, InterlockedCompareExchange((volatile long*)(Memory + addr), (u32)CPU.GPR[rs], reservation.data32) == reservation.data32);
+			reservation.clear();
+		}
+		else
+		{
+			CPU.SetCR_EQ(0, false);
+			if (lock.tid == reservation.owner) reservation.clear();
 		}
 	}
 	void STWX(u32 rs, u32 ra, u32 rb)
@ -2709,23 +2709,19 @@ private:
 	}
 	void STDCX_(u32 rs, u32 ra, u32 rb)
 	{
-		CPU.SetCR(0, CPU.XER.SO ? CR_SO : 0);
-		if(!CPU.reserve)
-		{
-			const u64 addr = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];
+		const u64 addr = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];

-			if(addr == CPU.reserve_addr)
-			{
-				Memory.Write64(addr, CPU.GPR[rs]);
-				CPU.SetCR_EQ(0, true);
-			}
-			else
-			{
-				static const bool u = 0;
-				if(u) Memory.Write64(addr, CPU.GPR[rs]);
-				CPU.SetCR_EQ(0, u);
-				CPU.reserve = false;
-			}
+		SMutexLocker lock(reservation.mutex);
+		if (lock.tid == reservation.owner && reservation.addr == addr && reservation.size == 8)
+		{
+			// Memory.Write64(addr, CPU.GPR[rs]);
+			CPU.SetCR_EQ(0, InterlockedCompareExchange64((volatile long long*)(Memory + addr), CPU.GPR[rs], reservation.data64) == reservation.data64);
+			reservation.clear();
+		}
+		else
+		{
+			CPU.SetCR_EQ(0, false);
+			if (lock.tid == reservation.owner) reservation.clear();
 		}
 	}
 	void STBX(u32 rs, u32 ra, u32 rb)
--- a/rpcs3/Emu/Cell/PPUThread.cpp
+++ b/rpcs3/Emu/Cell/PPUThread.cpp
@ -44,9 +44,6 @@ void PPUThread::DoReset()
 	VSCR.VSCR	= 0;

 	cycle = 0;
-
-	reserve = false;
-	reserve_addr = 0;
 }

 void PPUThread::AddArgv(const wxString& arg)
@ -67,14 +64,16 @@ void PPUThread::InitRegs()

 	SetPc(pc);

-	const s32 thread_num = Emu.GetCPU().GetThreadNumById(GetType(), GetId());
+	/*
+	const s32 thread_num = Emu.GetCPU().GetThread NumById(GetType(), GetId());

 	if(thread_num < 0)
 	{
-		ConLog.Error("GetThreadNumById failed.");
+		ConLog.Error("GetThread NumById failed.");
 		Emu.Pause();
 		return;
 	}
+	*/

 	/*
 	const s32 tls_size = Emu.GetTLSFilesz() * thread_num;
@ -119,7 +118,7 @@ void PPUThread::InitRegs()
 		GPR[6] = m_args[3];
 	}

-	u32 prx_mem = Memory.PRXMem.Alloc(0x10000);
+	u32 prx_mem = Memory.PRXMem.AllocAlign(0x10000);
 	Memory.Write64(prx_mem, 0xDEADBEEFABADCAFE);

 	GPR[0] = pc;
--- a/rpcs3/Emu/Cell/PPUThread.h
+++ b/rpcs3/Emu/Cell/PPUThread.h
@ -602,9 +602,6 @@ public:
 		};
 	};

-	u64 reserve_addr;
-	bool reserve;
-
 	u64 cycle;

 public:
--- a/rpcs3/Emu/Cell/SPUInterpreter.h
+++ b/rpcs3/Emu/Cell/SPUInterpreter.h
@ -32,15 +32,8 @@ private:
 	//0 - 10
 	void STOP(u32 code)
 	{
-		if(code & 0x2000)
-		{
-			CPU.SetExitStatus(code & 0xfff);
-		}
-		else
-		{
-			ConLog.Warning("STOP: 0x%x", code);
-			//Emu.Pause();
-		}
+		ConLog.Warning("STOP: 0x%x (m_exit_status -> 0)", code);
+		CPU.SetExitStatus(0);
 		CPU.Stop();
 	}
 	void LNOP()
@ -146,7 +139,7 @@ private:
 	void ROTH(u32 rt, u32 ra, u32 rb)
 	{
 		for (int h = 0; h < 8; h++) 
-			CPU.GPR[rt]._u16[h] = (CPU.GPR[ra]._u16[h] << (CPU.GPR[rb]._u16[h] & 0xf)) | (CPU.GPR[ra]._u16[h] >> (16 - (CPU.GPR[rb]._u32[h] & 0xf)));
+			CPU.GPR[rt]._u16[h] = (CPU.GPR[ra]._u16[h] << (CPU.GPR[rb]._u16[h] & 0xf)) | (CPU.GPR[ra]._u16[h] >> (16 - (CPU.GPR[rb]._u16[h] & 0xf)));
 	}
 	void ROTHM(u32 rt, u32 ra, u32 rb)
 	{
--- a/rpcs3/Emu/Cell/SPUThread.cpp
+++ b/rpcs3/Emu/Cell/SPUThread.cpp
@ -61,7 +61,7 @@ void SPUThread::InitRegs()

 u64 SPUThread::GetFreeStackSize() const
 {
-	return (GetStackAddr() + GetStackSize()) - GPR[1]._u64[3];
+	return (GetStackAddr() + GetStackSize()) - GPR[1]._u32[3];
 }

 void SPUThread::DoRun()
--- a/rpcs3/Emu/Cell/SPUThread.h
+++ b/rpcs3/Emu/Cell/SPUThread.h
@ -4,8 +4,6 @@
 #include "MFC.h"
 #include <mutex>

-extern std::mutex g_SyncMutex; //can provide compatability for CellSyncMutex through SPU<>PPU and SPU<>SPU
-
 static const char* spu_reg_name[128] =
 {
 	"$LR",  "$SP",  "$2",   "$3",   "$4",   "$5",   "$6",   "$7",
@ -391,7 +389,7 @@ public:
 			else
 			{
 #ifdef _M_X64
-				_InterlockedOr64((volatile __int64*)m_indval, ((u64)value << 32) | 1);
+				InterlockedOr64((volatile __int64*)m_indval, ((u64)value << 32) | 1);
 #else
 				ConLog.Error("PushUncond_OR(): no code compiled");
 #endif
@ -516,34 +514,80 @@ public:
 		case MFC_PUT_CMD:
 		case MFC_GET_CMD:
 		{
-			if (enable_log)	ConLog.Write("DMA %s%s%s: lsa = 0x%x, ea = 0x%llx, tag = 0x%x, size = 0x%x, cmd = 0x%x", 
+			/* if (enable_log) ConLog.Write("DMA %s%s%s: lsa = 0x%x, ea = 0x%llx, tag = 0x%x, size = 0x%x, cmd = 0x%x", 
 				op & MFC_PUT_CMD ? "PUT" : "GET",
 				op & MFC_BARRIER_MASK ? "B" : "", 
 				op & MFC_FENCE_MASK ? "F" : "", 
-				lsa, ea, tag, size, cmd);
-			MFCArgs.CMDStatus.SetValue(dmac.Cmd(cmd, tag, lsa, ea, size));
+				lsa, ea, tag, size, cmd); */
+			if (op & MFC_PUT_CMD)
+			{
+				SMutexLocker lock(reservation.mutex);
+				MFCArgs.CMDStatus.SetValue(dmac.Cmd(cmd, tag, lsa, ea, size));
+				if ((reservation.addr + reservation.size > ea && reservation.addr <= ea + size) || 
+					(ea + size > reservation.addr && ea <= reservation.addr + reservation.size))
+				{
+					reservation.clear();
+				}
+			}
+			else
+			{
+				MFCArgs.CMDStatus.SetValue(dmac.Cmd(cmd, tag, lsa, ea, size));
+			}
 		}
 		break;

 		case MFC_GETLLAR_CMD:
 		case MFC_PUTLLC_CMD:
 		case MFC_PUTLLUC_CMD:
+		case MFC_PUTQLLUC_CMD:
 		{
-			if (op == MFC_GETLLAR_CMD) 
-			{
-				g_SyncMutex.lock();
-			}
-			
-			ConLog.Warning("DMA %s: lsa=0x%x, ea = 0x%llx, (tag) = 0x%x, (size) = 0x%x, cmd = 0x%x",
-				op == MFC_GETLLAR_CMD ? "GETLLAR" : op == MFC_PUTLLC_CMD ? "PUTLLC" : "PUTLLUC",
+			if (enable_log) ConLog.Write("DMA %s: lsa=0x%x, ea = 0x%llx, (tag) = 0x%x, (size) = 0x%x, cmd = 0x%x",
+				op == MFC_GETLLAR_CMD ? "GETLLAR" : op == MFC_PUTLLC_CMD ? "PUTLLC" : op == MFC_PUTLLUC_CMD ? "PUTLLUC" : "PUTQLLUC",
 				lsa, ea, tag, size, cmd);

-			dmac.ProcessCmd(op == MFC_GETLLAR_CMD ? MFC_GET_CMD : MFC_PUT_CMD, tag, lsa, ea, 128);
-			Prxy.AtomicStat.PushUncond(op == MFC_GETLLAR_CMD ? MFC_GETLLAR_SUCCESS : op == MFC_PUTLLC_CMD ? MFC_PUTLLC_SUCCESS : MFC_PUTLLUC_SUCCESS);
-
-			if (op == MFC_PUTLLC_CMD || op == MFC_PUTLLUC_CMD) 
+			if (op == MFC_GETLLAR_CMD) // get reservation
 			{
-				g_SyncMutex.unlock();
+				SMutexLocker lock(reservation.mutex);
+				reservation.owner = lock.tid;
+				reservation.addr = ea;
+				reservation.size = 128;
+				dmac.ProcessCmd(MFC_GET_CMD, tag, lsa, ea, 128);
+				Prxy.AtomicStat.PushUncond(MFC_GETLLAR_SUCCESS);
+			}
+			else if (op == MFC_PUTLLC_CMD) // store conditional
+			{
+				SMutexLocker lock(reservation.mutex);
+				if (reservation.owner == lock.tid) // succeeded
+				{
+					if (reservation.addr == ea && reservation.size == 128)
+					{
+						dmac.ProcessCmd(MFC_PUT_CMD, tag, lsa, ea, 128);
+						Prxy.AtomicStat.PushUncond(MFC_PUTLLC_SUCCESS);
+					}
+					else
+					{
+						Prxy.AtomicStat.PushUncond(MFC_PUTLLC_FAILURE);
+					}
+					reservation.clear();
+				}
+				else // failed
+				{
+					Prxy.AtomicStat.PushUncond(MFC_PUTLLC_FAILURE);
+				}
+			}
+			else // store unconditional
+			{
+				SMutexLocker lock(reservation.mutex);
+				dmac.ProcessCmd(MFC_PUT_CMD, tag, lsa, ea, 128);
+				if (op == MFC_PUTLLUC_CMD)
+				{
+					Prxy.AtomicStat.PushUncond(MFC_PUTLLUC_SUCCESS);
+				}
+				if ((reservation.addr + reservation.size > ea && reservation.addr <= ea + size) || 
+					(ea + size > reservation.addr && ea <= reservation.addr + reservation.size))
+				{
+					reservation.clear();
+				}
 			}
 		}
 		break;
--- a/rpcs3/Emu/Memory/DynamicMemoryBlockBase.inl
+++ b/rpcs3/Emu/Memory/DynamicMemoryBlockBase.inl
@ -10,11 +10,13 @@ DynamicMemoryBlockBase<PT>::DynamicMemoryBlockBase()
 template<typename PT>
 const u32 DynamicMemoryBlockBase<PT>::GetUsedSize() const
 {
+	std::lock_guard<std::mutex> lock(m_lock);
+
 	u32 size = 0;

-	for(u32 i=0; i<m_used_mem.GetCount(); ++i)
+	for(u32 i=0; i<m_allocated.GetCount(); ++i)
 	{
-		size += m_used_mem[i].size;
+		size += m_allocated[i].size;
 	}

 	return size;
@ -35,38 +37,51 @@ bool DynamicMemoryBlockBase<PT>::IsInMyRange(const u64 addr, const u32 size)
 template<typename PT>
 bool DynamicMemoryBlockBase<PT>::IsMyAddress(const u64 addr)
 {
-	for(u32 i=0; i<m_used_mem.GetCount(); ++i)
-	{
-		if(addr >= m_used_mem[i].addr && addr < m_used_mem[i].addr + m_used_mem[i].size)
-		{
-			return true;
-		}
-	}
+	if (!IsInMyRange(addr)) return false;

-	return false;
+	const u32 index = MemoryBlock::FixAddr(addr) >> 12;
+
+	return m_pages[index] && !m_locked[index];
 }

 template<typename PT>
 MemoryBlock* DynamicMemoryBlockBase<PT>::SetRange(const u64 start, const u32 size)
 {
-	m_max_size = size;
+	std::lock_guard<std::mutex> lock(m_lock);
+
+	m_max_size = PAGE_4K(size);
 	MemoryBlock::SetRange(start, 0);

+	const u32 page_count = m_max_size >> 12;
+	m_pages.SetCount(page_count);
+	m_locked.SetCount(page_count);
+	memset(m_pages.GetPtr(), 0, sizeof(u8*) * page_count);
+	memset(m_locked.GetPtr(), 0, sizeof(u8*) * page_count);
+
 	return this;
 }

 template<typename PT>
 void DynamicMemoryBlockBase<PT>::Delete()
 {
-	m_used_mem.Clear();
+	std::lock_guard<std::mutex> lock(m_lock);
+
+	m_allocated.Clear();
 	m_max_size = 0;

+	m_pages.Clear();
+	m_locked.Clear();
+
 	MemoryBlock::Delete();
 }

 template<typename PT>
-bool DynamicMemoryBlockBase<PT>::Alloc(u64 addr, u32 size)
+bool DynamicMemoryBlockBase<PT>::AllocFixed(u64 addr, u32 size)
 {
+	size = PAGE_4K(size + (addr & 4095)); // align size
+
+	addr &= ~4095; // align start address
+
 	if(!IsInMyRange(addr, size))
 	{
 		assert(0);
@ -78,43 +93,77 @@ bool DynamicMemoryBlockBase<PT>::Alloc(u64 addr, u32 size)
 		return false;
 	}

-	for(u32 i=0; i<m_used_mem.GetCount(); ++i)
+	std::lock_guard<std::mutex> lock(m_lock);
+
+	for(u32 i=0; i<m_allocated.GetCount(); ++i)
 	{
-		if(addr >= m_used_mem[i].addr && addr < m_used_mem[i].addr + m_used_mem[i].size) return false;
+		if(addr >= m_allocated[i].addr && addr < m_allocated[i].addr + m_allocated[i].size) return false;
 	}

-	AppendUsedMem(addr, size);
+	AppendMem(addr, size);

 	return true;
 }

 template<typename PT>
-void DynamicMemoryBlockBase<PT>::AppendUsedMem(u64 addr, u32 size)
+void DynamicMemoryBlockBase<PT>::AppendMem(u64 addr, u32 size) /* private */
 {
-	m_used_mem.Move(new MemBlockInfo(addr, size));
+	u8* pointer = (u8*)m_allocated[m_allocated.Move(new MemBlockInfo(addr, size))].mem;
+
+	const u32 first = MemoryBlock::FixAddr(addr) >> 12;
+
+	const u32 last = first + ((size - 1) >> 12);
+
+	for (u32 i = first; i <= last; i++)
+	{
+		m_pages[i] = pointer;
+		m_locked[i] = nullptr;
+		pointer += 4096;
+	}
 }

 template<typename PT>
-u64 DynamicMemoryBlockBase<PT>::Alloc(u32 size)
+u64 DynamicMemoryBlockBase<PT>::AllocAlign(u32 size, u32 align)
 {
-	for(u64 addr = MemoryBlock::GetStartAddr(); addr <= MemoryBlock::GetEndAddr() - size;)
+	size = PAGE_4K(size);
+	u32 exsize;
+
+	if (align <= 4096)
+	{
+		align = 0;
+		exsize = size;
+	}
+	else
+	{
+		align &= ~4095;
+		exsize = size + align - 1;
+	}
+
+	std::lock_guard<std::mutex> lock(m_lock);
+
+	for(u64 addr = MemoryBlock::GetStartAddr(); addr <= MemoryBlock::GetEndAddr() - exsize;)
 	{
 		bool is_good_addr = true;

-		for(u32 i=0; i<m_used_mem.GetCount(); ++i)
+		for(u32 i=0; i<m_allocated.GetCount(); ++i)
 		{
-			if((addr >= m_used_mem[i].addr && addr < m_used_mem[i].addr + m_used_mem[i].size) ||
-				(m_used_mem[i].addr >= addr && m_used_mem[i].addr < addr + size))
+			if((addr >= m_allocated[i].addr && addr < m_allocated[i].addr + m_allocated[i].size) ||
+				(m_allocated[i].addr >= addr && m_allocated[i].addr < addr + exsize))
 			{
 				is_good_addr = false;
-				addr = m_used_mem[i].addr + m_used_mem[i].size;
+				addr = m_allocated[i].addr + m_allocated[i].size;
 				break;
 			}
 		}

 		if(!is_good_addr) continue;

-		AppendUsedMem(addr, size);
+		if (align)
+		{
+			addr = (addr + (align - 1)) & ~(align - 1);
+		}
+
+		AppendMem(addr, size);

 		return addr;
 	}
@ -125,18 +174,38 @@ u64 DynamicMemoryBlockBase<PT>::Alloc(u32 size)
 template<typename PT>
 bool DynamicMemoryBlockBase<PT>::Alloc()
 {
-	return Alloc(GetSize() - GetUsedSize()) != 0;
+	return AllocAlign(GetSize() - GetUsedSize()) != 0;
 }

 template<typename PT>
 bool DynamicMemoryBlockBase<PT>::Free(u64 addr)
-{	
-	for(u32 i=0; i<m_used_mem.GetCount(); ++i)
+{
+	std::lock_guard<std::mutex> lock(m_lock);
+
+	for (u32 num = 0; num < m_allocated.GetCount(); num++)
 	{
-		if(addr == m_used_mem[i].addr)
+		if (addr == m_allocated[num].addr)
 		{
-			if(IsLocked(m_used_mem[i].addr)) return false;
-			m_used_mem.RemoveAt(i);
+			/* if(IsLocked(m_allocated[num].addr)) return false; */
+
+			const u32 first = MemoryBlock::FixAddr(addr) >> 12;
+
+			const u32 last = first + ((m_allocated[num].size - 1) >> 12);
+
+			// check if locked:
+			for (u32 i = first; i <= last; i++)
+			{
+				if (!m_pages[i] || m_locked[i]) return false;
+			}
+
+			// clear pointers:
+			for (u32 i = first; i <= last; i++)
+			{
+				m_pages[i] = nullptr;
+				m_locked[i] = nullptr;
+			}
+
+			m_allocated.RemoveAt(num);
 			return true;
 		}
 	}
@ -145,15 +214,15 @@ bool DynamicMemoryBlockBase<PT>::Free(u64 addr)
 }

 template<typename PT>
-u8* DynamicMemoryBlockBase<PT>::GetMem(u64 addr) const
+u8* DynamicMemoryBlockBase<PT>::GetMem(u64 addr) const // lock-free, addr is fixed
 {
-	for(u32 i=0; i<m_used_mem.GetCount(); ++i)
-	{
-		u64 _addr = MemoryBlock::FixAddr(m_used_mem[i].addr);
+	const u32 index = addr >> 12;

-		if(addr >= _addr && addr < _addr + m_used_mem[i].size)
+	if (index < m_pages.GetCount())
+	{
+		if (u8* res = m_pages[index])
 		{
-			return (u8*)m_used_mem[i].mem + addr - _addr;
+			return res + (addr & 4095);
 		}
 	}

@ -163,28 +232,28 @@ u8* DynamicMemoryBlockBase<PT>::GetMem(u64 addr) const
 }

 template<typename PT>
-bool DynamicMemoryBlockBase<PT>::IsLocked(const u64 addr)
+bool DynamicMemoryBlockBase<PT>::IsLocked(u64 addr) // lock-free
 {
-	for(u32 i=0; i<m_locked_mem.GetCount(); ++i)
+	if (IsInMyRange(addr))
 	{
-		if(addr == m_locked_mem[i].addr)
+		const u32 index = MemoryBlock::FixAddr(addr) >> 12;
+
+		if (index < m_locked.GetCount())
 		{
-			return true;
+			if (m_locked[index]) return true;
 		}
 	}

 	return false;
 }

-template<typename PT>
-void DynamicMemoryBlockBase<PT>::AppendLockedMem(u64 addr, u32 size)
-{
-	m_locked_mem.Move(new MemBlockInfo(addr, size));
-}
-
 template<typename PT>
 bool DynamicMemoryBlockBase<PT>::Lock(u64 addr, u32 size)
 {
+	size = PAGE_4K(size); // align size
+
+	addr &= ~4095; // align start address
+
 	if(!IsInMyRange(addr, size))
 	{
 		assert(0);
@ -196,33 +265,58 @@ bool DynamicMemoryBlockBase<PT>::Lock(u64 addr, u32 size)
 		return false;
 	}

-	AppendLockedMem(addr, size);
+	const u32 first = MemoryBlock::FixAddr(addr) >> 12;
+
+	const u32 last = first + ((size - 1) >> 12);
+
+	for (u32 i = first; i <= last; i++)
+	{
+		if (u8* pointer = m_pages[i])
+		{
+			m_locked[i] = pointer;
+			m_pages[i] = nullptr;
+		}
+		else // already locked or empty
+		{
+		}
+	}

 	return true;
 }

 template<typename PT>
-bool DynamicMemoryBlockBase<PT>::Unlock(u64 addr , u32 size)
+bool DynamicMemoryBlockBase<PT>::Unlock(u64 addr, u32 size)
 {	
-	for(u32 i=0; i<m_locked_mem.GetCount(); ++i)
+	size = PAGE_4K(size); // align size
+
+	addr &= ~4095; // align start address
+
+	if(!IsInMyRange(addr, size))
 	{
-		if(addr == m_locked_mem[i].addr)
+		assert(0);
+		return false;
+	}
+
+	if(IsMyAddress(addr) || IsMyAddress(addr + size - 1))
+	{
+		return false;
+	}
+
+	const u32 first = MemoryBlock::FixAddr(addr) >> 12;
+
+	const u32 last = first + ((size - 1) >> 12);
+
+	for (u32 i = first; i <= last; i++)
+	{
+		if (u8* pointer = m_locked[i])
+		{
+			m_pages[i] = pointer;
+			m_locked[i] = nullptr;
+		}
+		else // already unlocked or empty
 		{
-			if(m_locked_mem.Get(i).size > size)
-			{
-				m_locked_mem.Get(i).size -= size;
-			}
-			else if(m_locked_mem.Get(i).size == size)
-			{
-				m_locked_mem.RemoveAt(i);
-			}
-			else
-			{
-				return false;
-			}
-			return true;
 		}
 	}

-	return false;
+	return true;
 }
--- a/rpcs3/Emu/Memory/Memory.h
+++ b/rpcs3/Emu/Memory/Memory.h
@ -341,7 +341,7 @@ public:

 	u64 Alloc(const u32 size, const u32 align)
 	{
-		return UserMemory->Alloc(AlignAddr(size, align));
+		return UserMemory->AllocAlign(size, align);
 	}

 	bool Free(const u64 addr)
--- a/rpcs3/Emu/Memory/MemoryBlock.h
+++ b/rpcs3/Emu/Memory/MemoryBlock.h
@ -1,5 +1,7 @@
 #pragma once

+#define PAGE_4K(x) (x + 4095) & ~(4095)
+
 struct MemInfo
 {
 	u64 addr;
@ -21,8 +23,8 @@ struct MemBlockInfo : public MemInfo
 	void* mem;

 	MemBlockInfo(u64 _addr, u32 _size)
-		: MemInfo(_addr, _size)
-		, mem(malloc(_size))
+		: MemInfo(_addr, PAGE_4K(_size))
+		, mem(_aligned_malloc(PAGE_4K(_size), 128))
 	{
 		if(!mem)
 		{
@ -35,7 +37,7 @@ struct MemBlockInfo : public MemInfo

 	~MemBlockInfo()
 	{
-		free(mem);
+		_aligned_free(mem);
 		mem = nullptr;
 	}
 };
@ -120,8 +122,8 @@ public:
 	u8* GetMem() const { return mem; }
 	virtual u8* GetMem(u64 addr) const { return mem + addr; }

-	virtual bool Alloc(u64 addr, u32 size) { return false; }
-	virtual u64 Alloc(u32 size) { return 0; }
+	virtual bool AllocFixed(u64 addr, u32 size) { return false; }
+	virtual u64 AllocAlign(u32 size, u32 align = 1) { return 0; }
 	virtual bool Alloc() { return false; }
 	virtual bool Free(u64 addr) { return false; }
 	virtual bool Lock(u64 addr, u32 size) { return false; }
@ -190,8 +192,11 @@ class NullMemoryBlock : public MemoryBlock
 template<typename PT>
 class DynamicMemoryBlockBase : public PT
 {
-	Array<MemBlockInfo> m_used_mem;
-	Array<MemBlockInfo> m_locked_mem;
+	mutable std::mutex m_lock;
+	Array<MemBlockInfo> m_allocated; // allocation info
+	Array<u8*> m_pages; // real addresses of every 4096 byte pages (array size should be fixed)
+	Array<u8*> m_locked; // locked pages should be moved here
+	
 	u32 m_max_size;

 public:
@ -209,8 +214,8 @@ public:

 	virtual void Delete();

-	virtual bool Alloc(u64 addr, u32 size);
-	virtual u64 Alloc(u32 size);
+	virtual bool AllocFixed(u64 addr, u32 size);
+	virtual u64 AllocAlign(u32 size, u32 align = 1);
 	virtual bool Alloc();
 	virtual bool Free(u64 addr);
 	virtual bool Lock(u64 addr, u32 size);
@ -219,8 +224,7 @@ public:
 	virtual u8* GetMem(u64 addr) const;

 private:
-	void AppendUsedMem(u64 addr, u32 size);
-	void AppendLockedMem(u64 addr, u32 size);
+	void AppendMem(u64 addr, u32 size);
 };

 class VirtualMemoryBlock : public MemoryBlock
--- a/rpcs3/Emu/SysCalls/Modules/cellAudio.cpp
+++ b/rpcs3/Emu/SysCalls/Modules/cellAudio.cpp
@ -83,7 +83,7 @@ struct CellAudioPortConfig
 struct CellAudioConfig  //custom structure
 {
 	bool m_is_audio_initialized;
-	bool m_is_audio_port_open;
+	bool m_is_audio_port_opened;
 	bool m_is_audio_port_started;
 };

@ -216,7 +216,7 @@ int cellAudioPortOpen(mem_ptr_t<CellAudioPortParam> audioParam, mem32_t portNum)
 	cellAudio.Warning("cellAudioPortOpen(audioParam_addr=0x%x,portNum_addr=0x%x)",audioParam.GetAddr(),portNum.GetAddr());

 	if(!audioParam.IsGood() || !portNum.IsGood()) return CELL_AUDIO_ERROR_PORT_OPEN;
-	m_config->m_is_audio_port_open = true;
+	m_config->m_is_audio_port_opened = true;
 	
 	m_param->nChannel = audioParam->nChannel;
 	m_param->nBlock = audioParam->nBlock;
@ -238,13 +238,13 @@ int cellAudioGetPortConfig(u32 portNum, mem_ptr_t<CellAudioPortConfig> portConfi

 	//if(portNum > 7) return CELL_AUDIO_ERROR_PORT_FULL;

-	if(m_config->m_is_audio_port_open == false)
+	if(!m_config->m_is_audio_port_opened)
 	{
 		portConfig->status = CELL_AUDIO_STATUS_CLOSE;
 		return CELL_OK;
 	}

-	if(m_config->m_is_audio_port_started == true)
+	if(m_config->m_is_audio_port_started)
 	{
 		portConfig->status = CELL_AUDIO_STATUS_RUN;
 	}
@ -268,7 +268,7 @@ int cellAudioPortStart(u32 portNum)
 {
 	cellAudio.Warning("cellAudioPortStart(portNum=0x%x)",portNum);

-	if (m_config->m_is_audio_port_open == true)
+	if (!m_config->m_is_audio_port_opened)
 		return CELL_AUDIO_ERROR_PORT_OPEN;
 	
 	m_config->m_is_audio_port_started = true;
@ -279,10 +279,10 @@ int cellAudioPortClose(u32 portNum)
 {
 	cellAudio.Warning("cellAudioPortClose(portNum=0x%x)",portNum);
 	
-	if (m_config->m_is_audio_port_open == false)
+	if (!m_config->m_is_audio_port_opened)
 		return CELL_AUDIO_ERROR_PORT_NOT_OPEN;
 	
-	m_config->m_is_audio_port_open = false;
+	m_config->m_is_audio_port_opened = false;
 	return CELL_OK;
 }

@ -1016,6 +1016,6 @@ void cellAudio_init()
 void cellAudio_unload()
 {
 	m_config->m_is_audio_initialized = false;
-	m_config->m_is_audio_port_open = false;
+	m_config->m_is_audio_port_opened = false;
 	m_config->m_is_audio_port_started = false;
 }
--- a/rpcs3/Emu/SysCalls/Modules/cellGcmSys.cpp
+++ b/rpcs3/Emu/SysCalls/Modules/cellGcmSys.cpp
@ -63,7 +63,7 @@ int cellGcmInit(u32 context_addr, u32 cmdSize, u32 ioSize, u32 ioAddress)
 	{
 		local_size = 0xf900000; //TODO
 		local_addr = Memory.RSXFBMem.GetStartAddr();
-		Memory.RSXFBMem.Alloc(local_size);
+		Memory.RSXFBMem.AllocAlign(local_size);
 	}

 	cellGcmSys.Warning("*** local memory(addr=0x%x, size=0x%x)", local_addr, local_size);
@ -78,7 +78,7 @@ int cellGcmInit(u32 context_addr, u32 cmdSize, u32 ioSize, u32 ioAddress)
 	current_config.coreFrequency = re32(500000000);

 	InitOffsetTable();
-	Memory.RSXCMDMem.Alloc(cmdSize);
+	Memory.RSXCMDMem.AllocAlign(cmdSize);
 	Memory.MemoryBlocks.push_back(Memory.RSXIOMem.SetRange(0x50000000, 0x10000000/*256MB*/));//TODO: implement allocateAdressSpace in memoryBase
 	cellGcmMapEaIoAddress(ioAddress, 0, ioSize);

@ -89,7 +89,7 @@ int cellGcmInit(u32 context_addr, u32 cmdSize, u32 ioSize, u32 ioAddress)
 	current_context.current = current_context.begin;
 	current_context.callback = re32(Emu.GetRSXCallback() - 4);

-	gcm_info.context_addr = Memory.MainMem.Alloc(0x1000);
+	gcm_info.context_addr = Memory.MainMem.AllocAlign(0x1000);
 	gcm_info.control_addr = gcm_info.context_addr + 0x40;

 	Memory.WriteData(gcm_info.context_addr, current_context);
@ -167,7 +167,7 @@ int cellGcmSetPrepareFlip(mem_ptr_t<CellGcmContextData> ctxt, u32 id)
 		return CELL_GCM_ERROR_FAILURE;
 	}

-	GSLockCurrent gslock(GS_LOCK_WAIT_FLUSH);
+	GSLockCurrent gslock(GS_LOCK_WAIT_FLUSH); // could stall on exit

 	u32 current = re(ctxt->current);
 	u32 end = re(ctxt->end);
@ -682,7 +682,7 @@ int32_t cellGcmMapLocalMemory(u64 address, u64 size)
 	{
 		local_size = 0xf900000; //TODO
 		local_addr = Memory.RSXFBMem.GetStartAddr();
-		Memory.RSXFBMem.Alloc(local_size);
+		Memory.RSXFBMem.AllocAlign(local_size);
 		Memory.Write32(address, local_addr);
 		Memory.Write32(size, local_size);
 	}
@ -749,7 +749,7 @@ int32_t cellGcmUnmapEaIoAddress(u64 ea)
 		ea = ea >> 20;
 		io = Memory.Read16(offsetTable.io + (ea*sizeof(u16)));

-		for(int i=0; i<size; i++)
+		for(u32 i=0; i<size; i++)
 		{
 			Memory.Write16(offsetTable.io + ((ea+i)*sizeof(u16)), 0xFFFF);
 			Memory.Write16(offsetTable.ea + ((io+i)*sizeof(u16)), 0xFFFF);
@ -772,7 +772,7 @@ int32_t cellGcmUnmapIoAddress(u64 io)
 		io = io >> 20;
 		ea = Memory.Read16(offsetTable.ea + (io*sizeof(u16)));

-		for(int i=0; i<size; i++)
+		for(u32 i=0; i<size; i++)
 		{
 			Memory.Write16(offsetTable.io + ((ea+i)*sizeof(u16)), 0xFFFF);
 			Memory.Write16(offsetTable.ea + ((io+i)*sizeof(u16)), 0xFFFF);
--- a/rpcs3/Emu/SysCalls/Modules/cellPamf.cpp
+++ b/rpcs3/Emu/SysCalls/Modules/cellPamf.cpp
@ -480,7 +480,7 @@ int cellPamfReaderGetStreamInfo(mem_ptr_t<CellPamfReader> pSelf, u32 pInfo_addr,
 			pInfo->numberOfChannels = pAudio->channels;
 			pInfo->samplingFrequency = CELL_PAMF_FS_48kHz;

-			if (pAudio->bps = 0x40)
+			if (pAudio->bps == 0x40)
 				pInfo->bitsPerSample = CELL_PAMF_BIT_LENGTH_16;
 			else
 				//TODO: CELL_PAMF_BIT_LENGTH_24
--- a/rpcs3/Emu/SysCalls/Modules/cellSync.cpp
+++ b/rpcs3/Emu/SysCalls/Modules/cellSync.cpp
@ -4,9 +4,7 @@
 #include <mutex>

 void cellSync_init();
-void cellSync_unload();
-Module cellSync("cellSync", cellSync_init, nullptr, cellSync_unload);
-std::mutex g_SyncMutex;
+Module cellSync("cellSync", cellSync_init);

 // Return Codes
 enum
@ -56,8 +54,13 @@ int cellSyncMutexInitialize(mem_ptr_t<CellSyncMutex> mutex)
 		return CELL_SYNC_ERROR_ALIGN;
 	}

-	{ // global mutex
-		std::lock_guard<std::mutex> lock(g_SyncMutex); //???
+	{
+		SMutexLocker lock(reservation.mutex);
+		if ((reservation.addr + reservation.size > mutex.GetAddr() && reservation.addr <= mutex.GetAddr() + 4) || 
+			(mutex.GetAddr() + 4 > reservation.addr && mutex.GetAddr() <= reservation.addr + reservation.size))
+		{
+			reservation.clear();
+		}
 		mutex->m_data = 0;
 		return CELL_OK;
 	}
@ -77,8 +80,13 @@ int cellSyncMutexLock(mem_ptr_t<CellSyncMutex> mutex)
 	}

 	be_t<u16> old_order;
-	{ // global mutex
-		std::lock_guard<std::mutex> lock(g_SyncMutex);
+	{
+		SMutexLocker lock(reservation.mutex);
+		if ((reservation.addr + reservation.size > mutex.GetAddr() && reservation.addr <= mutex.GetAddr() + 4) || 
+			(mutex.GetAddr() + 4 > reservation.addr && mutex.GetAddr() <= reservation.addr + reservation.size))
+		{
+			reservation.clear();
+		}
 		old_order = mutex->m_order;
 		mutex->m_order = mutex->m_order + 1;
 		if (old_order == mutex->m_freed)
@ -98,7 +106,6 @@ int cellSyncMutexLock(mem_ptr_t<CellSyncMutex> mutex)
 				mutex.GetAddr(), (u16)old_order, (u16)mutex->m_order, (u16)mutex->m_freed);
 		}
 	}
-	//while (_InterlockedExchange((volatile long*)&mutex->m_data, 1)) Sleep(1);
 	_mm_mfence();
 	return CELL_OK;
 }
@ -115,8 +122,13 @@ int cellSyncMutexTryLock(mem_ptr_t<CellSyncMutex> mutex)
 	{
 		return CELL_SYNC_ERROR_ALIGN;
 	}
-	{ /* global mutex */
-		std::lock_guard<std::mutex> lock(g_SyncMutex);
+	{
+		SMutexLocker lock(reservation.mutex);
+		if ((reservation.addr + reservation.size > mutex.GetAddr() && reservation.addr <= mutex.GetAddr() + 4) || 
+			(mutex.GetAddr() + 4 > reservation.addr && mutex.GetAddr() <= reservation.addr + reservation.size))
+		{
+			reservation.clear();
+		}
 		if (mutex->m_order != mutex->m_freed)
 		{
 			return CELL_SYNC_ERROR_BUSY;
@ -140,7 +152,12 @@ int cellSyncMutexUnlock(mem_ptr_t<CellSyncMutex> mutex)
 	}

 	{ /* global mutex */
-		std::lock_guard<std::mutex> lock(g_SyncMutex);
+		SMutexLocker lock(reservation.mutex);
+		if ((reservation.addr + reservation.size > mutex.GetAddr() && reservation.addr <= mutex.GetAddr() + 4) || 
+			(mutex.GetAddr() + 4 > reservation.addr && mutex.GetAddr() <= reservation.addr + reservation.size))
+		{
+			reservation.clear();
+		}
 		mutex->m_freed = mutex->m_freed + 1;
 		return CELL_OK;
 	}
@ -152,9 +169,4 @@ void cellSync_init()
 	cellSync.AddFunc(0x1bb675c2, cellSyncMutexLock);
 	cellSync.AddFunc(0xd06918c4, cellSyncMutexTryLock);
 	cellSync.AddFunc(0x91f2b7b0, cellSyncMutexUnlock);
-}
-
-void cellSync_unload()
-{
-	g_SyncMutex.unlock();
 }
--- a/rpcs3/Emu/SysCalls/Modules/cellSysutil.cpp
+++ b/rpcs3/Emu/SysCalls/Modules/cellSysutil.cpp
@ -460,6 +460,11 @@ int cellSysutilCheckCallback()
 	cellSysutil.Log("cellSysutilCheckCallback()");
 	Emu.GetCallbackManager().m_exit_callback.Check();

+	CPUThread& thr = Emu.GetCallbackThread();
+
+	while(Emu.IsRunning() && thr.IsAlive())
+		Sleep(1);
+
 	return CELL_OK;
 }

--- a/rpcs3/Emu/SysCalls/Modules/sysPrxForUser.cpp
+++ b/rpcs3/Emu/SysCalls/Modules/sysPrxForUser.cpp
@ -164,6 +164,7 @@ void sysPrxForUser_init()
 	//sysPrxForUser.AddFunc(0xaede4b03, sys_heap_free);
 	//sysPrxForUser.AddFunc(0x8a561d92, sys_heap_delete_heap);
 	sysPrxForUser.AddFunc(0xb2fcf2c8, sys_heap_create_heap);
+	sysPrxForUser.AddFunc(0x44265c08, _sys_heap_memalign);

 	sysPrxForUser.AddFunc(0xb257540b, sys_mmapper_allocate_memory);
 	sysPrxForUser.AddFunc(0xdc578057, sys_mmapper_map_memory);
--- a/rpcs3/Emu/SysCalls/Modules/sys_fs.cpp
+++ b/rpcs3/Emu/SysCalls/Modules/sys_fs.cpp
@ -176,8 +176,8 @@ void fsAioRead(u32 fd, mem_ptr_t<CellFsAio> aio, int xid, mem_func_ptr_t<void (*
 	}

 	//start callback thread
-	//if(func)
-		//func.async(aio, error, xid, res);
+	if(func)
+		func.async(aio, error, xid, res);

 	ConLog.Warning("*** fsAioRead(fd=%d, offset=0x%llx, buf_addr=0x%x, size=%d, res=%d, xid=%d [%s])",
 		fd, (u64)aio->offset, buf_addr, (u64)aio->size, res, xid, path.c_str());
@ -195,6 +195,7 @@ int cellFsAioRead(mem_ptr_t<CellFsAio> aio, mem32_t aio_id, mem_func_ptr_t<void

 	thread t("fsAioRead", std::bind(fsAioRead, fd, aio, xid, func));
 	t.detach();
+	//fsAioRead(fd, aio, xid, func);

 	aio_id = xid;

--- a/rpcs3/Emu/SysCalls/SysCalls.cpp
+++ b/rpcs3/Emu/SysCalls/SysCalls.cpp
@ -87,20 +87,20 @@ static func_caller* sc_table[1024] =
 	bind_func(sys_rwlock_wunlock),							//127 (0x07F)
 	bind_func(sys_event_queue_create),						//128 (0x080)
 	null_func,												//129 (0x081)
-	bind_func(sys_event_queue_receive), null_func, bind_func(sys_event_flag_cancel), null_func, bind_func(sys_event_port_create), //134
+	bind_func(sys_event_queue_receive), null_func, bind_func(sys_event_flag_cancel), bind_func(sys_event_queue_drain), bind_func(sys_event_port_create), //134
 	null_func, bind_func(sys_event_port_connect_local), null_func, bind_func(sys_event_port_send), bind_func(sys_event_flag_get), //139
 	null_func, bind_func(sys_timer_usleep), bind_func(sys_timer_sleep), null_func, bind_func(sys_time_get_timezone), //144
 	bind_func(sys_time_get_current_time), bind_func(sys_time_get_system_time), bind_func(sys_time_get_timebase_frequency), null_func, null_func, //149
 	null_func, null_func, null_func, null_func, null_func, //154
 	null_func, bind_func(sys_spu_image_open), null_func, null_func, null_func, //159
 	bind_func(sys_raw_spu_create), null_func, null_func, null_func, null_func, //164
-	null_func, null_func, null_func, null_func, bind_func(sys_spu_initialize), //169
-	bind_func(sys_spu_thread_group_create), bind_func(sys_spu_thread_set_argument), bind_func(sys_spu_thread_initialize), //172
+	bind_func(sys_spu_thread_get_exit_status), bind_func(sys_spu_thread_set_argument), null_func, null_func, bind_func(sys_spu_initialize), //169
+	bind_func(sys_spu_thread_group_create), bind_func(sys_spu_thread_group_destroy), bind_func(sys_spu_thread_initialize), //172
 	bind_func(sys_spu_thread_group_start), bind_func(sys_spu_thread_group_suspend), //174
 	null_func, null_func, null_func, bind_func(sys_spu_thread_group_join), null_func, //179
 	null_func, bind_func(sys_spu_thread_write_ls), bind_func(sys_spu_thread_read_ls), null_func, bind_func(sys_spu_thread_write_snr), //184
 	null_func, null_func, bind_func(sys_spu_thread_set_spu_cfg), bind_func(sys_spu_thread_get_spu_cfg), null_func, //189
-	bind_func(sys_spu_thread_write_spu_mb), bind_func(sys_spu_thread_connect_event), null_func, null_func, null_func, //194
+	bind_func(sys_spu_thread_write_spu_mb), bind_func(sys_spu_thread_connect_event), null_func, bind_func(sys_spu_thread_bind_queue), null_func, //194
 	null_func, null_func, null_func, null_func, null_func, //199
 	null_func, null_func, null_func, null_func, null_func, //204
 	null_func, null_func, null_func, null_func, null_func, //209
--- a/rpcs3/Emu/SysCalls/SysCalls.h
+++ b/rpcs3/Emu/SysCalls/SysCalls.h
@ -7,6 +7,7 @@
 #include "lv2/SC_SPU_Thread.h"
 #include "lv2/SC_Lwmutex.h"
 #include "lv2/SC_Lwcond.h"
+#include "lv2/SC_Event_flag.h"
 #include "Emu/event.h"
 //#define SYSCALLS_DEBUG

@ -126,19 +127,22 @@ extern int sys_game_process_exitspawn(u32 path_addr, u32 argv_addr, u32 envp_add
 								u32 data, u32 data_size, int prio, u64 flags );

 //sys_event
-extern int sys_event_flag_create(u32 eflag_id_addr, u32 attr_addr, u64 init);
-extern int sys_event_flag_destroy(u32 eflag_id);
-extern int sys_event_flag_wait(u32 eflag_id, u64 bitptn, u32 mode, u32 result_addr, u32 timeout);
-extern int sys_event_flag_trywait(u32 eflag_id, u64 bitptn, u32 mode, u32 result_addr);
-extern int sys_event_flag_set(u32 eflag_id, u64 bitptn);
-extern int sys_event_flag_clear(u32 eflag_id, u64 bitptn);
-extern int sys_event_flag_cancel(u32 eflag_id, u32 num_addr);
-extern int sys_event_flag_get(u32 eflag_id, u32 flag_addr);
-extern int sys_event_queue_create(u32 equeue_id_addr, u32 attr_addr, u64 event_queue_key, int size);
+extern int sys_event_queue_create(mem32_t equeue_id, mem_ptr_t<sys_event_queue_attr> attr, u64 event_queue_key, int size);
 extern int sys_event_queue_receive(u32 equeue_id, u32 event_addr, u32 timeout);
 extern int sys_event_port_create(u32 eport_id_addr, int port_type, u64 name);
 extern int sys_event_port_connect_local(u32 event_port_id, u32 event_queue_id);
 extern int sys_event_port_send(u32 event_port_id, u64 data1, u64 data2, u64 data3);
+extern int sys_event_queue_drain(u32 event_queue_id);
+
+//sys_event_flag
+extern int sys_event_flag_create(mem32_t eflag_id, mem_ptr_t<sys_event_flag_attr> attr, u64 init);
+extern int sys_event_flag_destroy(u32 eflag_id);
+extern int sys_event_flag_wait(u32 eflag_id, u64 bitptn, u32 mode, mem64_t result, u32 timeout);
+extern int sys_event_flag_trywait(u32 eflag_id, u64 bitptn, u32 mode, mem64_t result);
+extern int sys_event_flag_set(u32 eflag_id, u64 bitptn);
+extern int sys_event_flag_clear(u32 eflag_id, u64 bitptn);
+extern int sys_event_flag_cancel(u32 eflag_id, mem32_t num);
+extern int sys_event_flag_get(u32 eflag_id, mem64_t flags);

 //sys_semaphore
 extern int sys_semaphore_create(u32 sem_addr, u32 attr_addr, int initial_val, int max_val);
@ -188,10 +192,9 @@ extern int sys_ppu_thread_get_priority(u32 thread_id, u32 prio_addr);
 extern int sys_ppu_thread_get_stack_information(u32 info_addr);
 extern int sys_ppu_thread_stop(u32 thread_id);
 extern int sys_ppu_thread_restart(u32 thread_id);
-extern int sys_ppu_thread_create(u32 thread_id_addr, u32 entry, u32 arg, int prio, u32 stacksize, u64 flags, u32 threadname_addr);
+extern int sys_ppu_thread_create(u32 thread_id_addr, u32 entry, u64 arg, int prio, u32 stacksize, u64 flags, u32 threadname_addr);
 extern void sys_ppu_thread_once(u32 once_ctrl_addr, u32 entry);
 extern int sys_ppu_thread_get_id(const u32 id_addr);
-extern int sys_spu_thread_group_connect_event_all_threads(u32 id, u32 eq, u64 req, u32 spup_addr);

 //memory
 extern int sys_memory_container_create(u32 cid_addr, u32 yield_size);
@ -293,17 +296,20 @@ extern int sys_tty_write(u32 ch, u64 buf_addr, u32 len, u64 pwritelen_addr);
 //sys_heap
 extern int sys_heap_create_heap(const u32 heap_addr, const u32 start_addr, const u32 size);
 extern int sys_heap_malloc(const u32 heap_addr, const u32 size);
+extern int _sys_heap_memalign(u32 heap_id, u32 align, u32 size, u64 p4);

 //sys_spu
 extern int sys_spu_image_open(mem_ptr_t<sys_spu_image> img, u32 path_addr);
 extern int sys_spu_thread_initialize(mem32_t thread, u32 group, u32 spu_num, mem_ptr_t<sys_spu_image> img, mem_ptr_t<sys_spu_thread_attribute> attr, mem_ptr_t<sys_spu_thread_argument> arg);
 extern int sys_spu_thread_set_argument(u32 id, mem_ptr_t<sys_spu_thread_argument> arg);
+extern int sys_spu_thread_group_destroy(u32 id);
 extern int sys_spu_thread_group_start(u32 id);
 extern int sys_spu_thread_group_suspend(u32 id);
 extern int sys_spu_thread_group_create(mem32_t id, u32 num, int prio, mem_ptr_t<sys_spu_thread_group_attribute> attr);
 extern int sys_spu_thread_create(mem32_t thread_id, mem32_t entry, u64 arg, int prio, u32 stacksize, u64 flags, u32 threadname_addr);
 extern int sys_spu_thread_connect_event(u32 id, u32 eq, u32 et, u8 spup);
 extern int sys_spu_thread_group_join(u32 id, mem32_t cause, mem32_t status);
+extern int sys_spu_thread_group_connect_event_all_threads(u32 id, u32 eq, u64 req, u32 spup_addr);
 extern int sys_raw_spu_create(mem32_t id, u32 attr_addr);
 extern int sys_spu_initialize(u32 max_usable_spu, u32 max_raw_spu);
 extern int sys_spu_thread_write_ls(u32 id, u32 address, u64 value, u32 type);
@ -312,6 +318,8 @@ extern int sys_spu_thread_write_spu_mb(u32 id, u32 value);
 extern int sys_spu_thread_set_spu_cfg(u32 id, u64 value);
 extern int sys_spu_thread_get_spu_cfg(u32 id, mem64_t value);
 extern int sys_spu_thread_write_snr(u32 id, u32 number, u32 value);
+extern int sys_spu_thread_bind_queue(u32 id, u32 spuq, u32 spuq_num);
+extern int sys_spu_thread_get_exit_status(u32 id, mem32_t status);

 //sys_time
 extern int sys_time_get_timezone(mem32_t timezone, mem32_t summertime);
--- a/rpcs3/Emu/SysCalls/lv2/SC_Condition.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Condition.cpp
@ -27,7 +27,7 @@ struct condition

 int sys_cond_create(u32 cond_addr, u32 mutex_id, u32 attr_addr)
 {
-	sys_cond.Log("sys_cond_create(cond_addr=0x%x, mutex_id=0x%x, attr_addr=%d)",
+	sys_cond.Warning("sys_cond_create(cond_addr=0x%x, mutex_id=0x%x, attr_addr=%d)",
 		cond_addr, mutex_id, attr_addr);

 	if(!Memory.IsGoodAddr(cond_addr) || !Memory.IsGoodAddr(attr_addr)) return CELL_EFAULT;
@ -68,9 +68,32 @@ int sys_cond_wait(u32 cond_id, u64 timeout)
 	condition* cond_data = nullptr;
 	if(!sys_cond.CheckId(cond_id, cond_data)) return CELL_ESRCH;

-	cond_data->cond.WaitTimeout(timeout ? timeout : INFINITE);
+	u32 counter = 0;
+	const u32 max_counter = timeout ? (timeout / 1000) : 20000;
+	do
+	{
+		if (Emu.IsStopped()) return CELL_ETIMEDOUT;

-	return CELL_OK;
+		switch (cond_data->cond.WaitTimeout(1))
+		{
+		case wxCOND_NO_ERROR: return CELL_OK;
+		case wxCOND_TIMEOUT: break;
+		default: return CELL_EPERM;
+		}
+
+		if (counter++ > max_counter)
+		{
+			if (!timeout) 
+			{
+				sys_cond.Warning("sys_cond_wait(cond_id=0x%x, timeout=0x%llx): TIMEOUT", cond_id, timeout);
+				counter = 0;
+			}
+			else
+			{
+				return CELL_ETIMEDOUT;
+			}
+		}		
+	} while (true);
 }

 int sys_cond_signal(u32 cond_id)
--- a/rpcs3/Emu/SysCalls/lv2/SC_Event.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Event.cpp
@ -5,128 +5,25 @@

 SysCallBase sys_event("sys_event");

-int sys_event_flag_create(u32 eflag_id_addr, u32 attr_addr, u64 init)
-{
-	sys_event.Warning("sys_event_flag_create(eflag_id_addr=0x%x, attr_addr=0x%x, init=0x%llx)", eflag_id_addr, attr_addr, init);
-
-	if(!Memory.IsGoodAddr(eflag_id_addr, 4) || !Memory.IsGoodAddr(attr_addr, sizeof(sys_event_flag_attr)))
-	{
-		return CELL_EFAULT;
-	}
-
-	sys_event_flag_attr attr = (sys_event_flag_attr&)Memory[attr_addr];
-	attr.protocol = re(attr.protocol);
-	attr.pshared = re(attr.pshared);
-	attr.ipc_key = re(attr.ipc_key);
-	attr.flags = re(attr.flags);
-	attr.type = re(attr.type);
-
-	sys_event.Warning("name = %s", attr.name);
-	sys_event.Warning("type = %d", attr.type);
-
-	Memory.Write32(eflag_id_addr, sys_event.GetNewId(new event_flag(init, attr)));
-
-	return CELL_OK;
-}
-
-int sys_event_flag_destroy(u32 eflag_id)
-{
-	sys_event.Warning("sys_event_flag_destroy(eflag_id=0x%x)", eflag_id);
-
-	if(!sys_event.CheckId(eflag_id)) return CELL_ESRCH;
-
-	Emu.GetIdManager().RemoveID(eflag_id);
-
-	return CELL_OK;
-}
-
-int sys_event_flag_wait(u32 eflag_id, u64 bitptn, u32 mode, u32 result_addr, u32 timeout)
-{
-	sys_event.Warning("Unimplemented function: sys_event_flag_wait(eflag_id=0x%x, bitptn=0x%llx, mode=0x%x, result_addr=0x%x, timeout=0x%x)"
-		, eflag_id, bitptn, mode, result_addr, timeout);
-	return CELL_OK;
-}
-
-int sys_event_flag_trywait(u32 eflag_id, u64 bitptn, u32 mode, u32 result_addr)
-{
-	sys_event.Warning("Unimplemented function: sys_event_flag_trywait(eflag_id=0x%x, bitptn=0x%llx, mode=0x%x, result_addr=0x%x)"
-		, eflag_id, bitptn, mode, result_addr);
-	return CELL_OK;
-}
-
-int sys_event_flag_set(u32 eflag_id, u64 bitptn)
-{
-	sys_event.Warning("sys_event_flag_set(eflag_id=0x%x, bitptn=0x%llx)", eflag_id, bitptn);
-
-	event_flag* event_flag_data = nullptr;
-	if(!sys_event.CheckId(eflag_id, event_flag_data)) return CELL_ESRCH;
-
-	event_flag_data->pattern |= bitptn;
-
-	return CELL_OK;
-}
-
-int sys_event_flag_clear(u32 eflag_id, u64 bitptn)
-{
-	sys_event.Warning("sys_event_flag_clear(eflag_id=0x%x, bitptn=0x%llx)", eflag_id, bitptn);
-
-	event_flag* event_flag_data = nullptr;
-	if(!sys_event.CheckId(eflag_id, event_flag_data)) return CELL_ESRCH;
-
-	event_flag_data->pattern &= bitptn;
-
-	return CELL_OK;
-}
-
-int sys_event_flag_cancel(u32 eflag_id, u32 num_addr)
-{
-	sys_event.Warning("Unimplemented function: sys_event_flag_cancel(eflag_id=0x%x, num_addr=0x%x)"
-		, eflag_id, num_addr);
-	return CELL_OK;
-}
-
-int sys_event_flag_get(u32 eflag_id, u32 flag_addr)
-{
-	sys_event.Warning("sys_event_flag_get(eflag_id=0x%x, flag_addr=0x%x)", eflag_id, flag_addr);
-	
-	if(!Memory.IsGoodAddr(flag_addr, 4))
-	{
-		return CELL_EFAULT;
-	}
-	
-	event_flag* event_flag_data = nullptr;
-	if(!sys_event.CheckId(eflag_id, event_flag_data)) return CELL_ESRCH;
-
-	Memory.Write64(flag_addr, event_flag_data->pattern);
-
-	return CELL_OK;
-}
-
 //128
-int sys_event_queue_create(u32 equeue_id_addr, u32 attr_addr, u64 event_queue_key, int size)
+int sys_event_queue_create(mem32_t equeue_id, mem_ptr_t<sys_event_queue_attr> attr, u64 event_queue_key, int size)
 {
 	sys_event.Warning("sys_event_queue_create(equeue_id_addr=0x%x, attr_addr=0x%x, event_queue_key=0x%llx, size=%d)",
-		equeue_id_addr, attr_addr, event_queue_key, size);
+		equeue_id.GetAddr(), attr.GetAddr(), event_queue_key, size);

 	if(size <= 0 || size > 127)
 	{
 		return CELL_EINVAL;
 	}

-	if(!Memory.IsGoodAddr(equeue_id_addr, 4) || !Memory.IsGoodAddr(attr_addr, sizeof(sys_event_queue_attr)))
+	if(!equeue_id.IsGood() || !attr.IsGood())
 	{
 		return CELL_EFAULT;
 	}

-	auto& attr = (sys_event_queue_attr&)Memory[attr_addr];
-	sys_event.Warning("name = %s", attr.name);
-	sys_event.Warning("type = %d", re(attr.type));
-	EventQueue* equeue = new EventQueue();
-	equeue->size = size;
-	equeue->pos = 0;
-	equeue->type = re(attr.type);
-	strncpy(equeue->name, attr.name, 8);
-	Memory.Write32(equeue_id_addr, sys_event.GetNewId(equeue));
+	equeue_id = sys_event.GetNewId(new EventQueue((u32)attr->protocol, (int)attr->type, attr->name_u64, event_queue_key, size));
+	sys_event.Warning("*** event_queue created[%s] (protocol=0x%x, type=0x%x): id = %d",
+		attr->name, (u32)attr->protocol, (int)attr->type, equeue_id.GetValue());

 	return CELL_OK;
 }
@ -176,10 +73,10 @@ int sys_event_queue_receive(u32 equeue_id, u32 event_addr, u32 timeout)
 			{
 				auto dst = (sys_event_data&)Memory[event_addr];

-				re(dst.source, equeue->ports[i]->name);
-				re(dst.data1, equeue->ports[i]->data1);
-				re(dst.data2, equeue->ports[i]->data2);
-				re(dst.data3, equeue->ports[i]->data3);
+				dst.source = equeue->ports[i]->name;
+				dst.data1 = equeue->ports[i]->data1;
+				dst.data2 = equeue->ports[i]->data2;
+				dst.data3 = equeue->ports[i]->data3;

 				equeue->ports[i]->has_data = false;

@ -264,5 +161,12 @@ int sys_event_port_send(u32 event_port_id, u64 data1, u64 data2, u64 data3)
 	eport->data2 = data2;
 	eport->data3 = data3;

+	return CELL_OK;
+}
+
+int sys_event_queue_drain(u32 event_queue_id)
+{
+	sys_event.Error("sys_event_queue_drain(event_queue_id=0x%x)", event_queue_id);
+
 	return CELL_OK;
 }
--- a/rpcs3/Emu/SysCalls/lv2/SC_Event_flag.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Event_flag.cpp
@ -0,0 +1,115 @@
+#include "stdafx.h"
+#include "Emu/SysCalls/SysCalls.h"
+#include "Emu/SysCalls/lv2/SC_Event_flag.h"
+
+SysCallBase sys_event_flag("sys_event_flag");
+
+int sys_event_flag_create(mem32_t eflag_id, mem_ptr_t<sys_event_flag_attr> attr, u64 init)
+{
+	sys_event_flag.Warning("sys_event_flag_create(eflag_id_addr=0x%x, attr_addr=0x%x, init=0x%llx)", eflag_id.GetAddr(), attr.GetAddr(), init);
+
+	if(!eflag_id.IsGood() || !attr.IsGood())
+	{
+		return CELL_EFAULT;
+	}
+
+	switch (attr->protocol.ToBE())
+	{
+	case se32(SYS_SYNC_PRIORITY): sys_event_flag.Warning("TODO: SYS_SYNC_PRIORITY attr"); break;
+	case se32(SYS_SYNC_RETRY): break;
+	case se32(SYS_SYNC_PRIORITY_INHERIT): sys_event_flag.Warning("TODO: SYS_SYNC_PRIORITY_INHERIT attr"); break;
+	case se32(SYS_SYNC_FIFO): sys_event_flag.Warning("TODO: SYS_SYNC_FIFO attr"); break;
+	default: return CELL_EINVAL;
+	}
+
+	if (attr->pshared.ToBE() != se32(0x200))
+	{
+		return CELL_EINVAL;
+	}
+
+	switch (attr->type.ToBE())
+	{
+	case se32(SYS_SYNC_WAITER_SINGLE): sys_event_flag.Warning("TODO: SYS_SYNC_WAITER_SINGLE type"); break;
+	case se32(SYS_SYNC_WAITER_MULTIPLE): sys_event_flag.Warning("TODO: SYS_SYNC_WAITER_MULTIPLE type"); break;
+	default: return CELL_EINVAL;
+	}
+
+	eflag_id = sys_event_flag.GetNewId(new event_flag(init, (u32)attr->protocol, (int)attr->type));
+
+	sys_event_flag.Warning("*** event_flag created[%s] (protocol=%d, type=%d): id = %d", attr->name, (u32)attr->protocol, (int)attr->type, eflag_id.GetValue());
+
+	return CELL_OK;
+}
+
+int sys_event_flag_destroy(u32 eflag_id)
+{
+	sys_event_flag.Warning("sys_event_flag_destroy(eflag_id=0x%x)", eflag_id);
+
+	event_flag* ef;
+	if(!sys_event_flag.CheckId(eflag_id, ef)) return CELL_ESRCH;
+
+	Emu.GetIdManager().RemoveID(eflag_id);
+
+	return CELL_OK;
+}
+
+int sys_event_flag_wait(u32 eflag_id, u64 bitptn, u32 mode, mem64_t result, u32 timeout)
+{
+	sys_event_flag.Error("sys_event_flag_wait(eflag_id=0x%x, bitptn=0x%llx, mode=0x%x, result_addr=0x%x, timeout=0x%x)",
+		eflag_id, bitptn, mode, result.GetAddr(), timeout);
+	return CELL_OK;
+}
+
+int sys_event_flag_trywait(u32 eflag_id, u64 bitptn, u32 mode, mem64_t result)
+{
+	sys_event_flag.Error("sys_event_flag_trywait(eflag_id=0x%x, bitptn=0x%llx, mode=0x%x, result_addr=0x%x)",
+		eflag_id, bitptn, mode, result.GetAddr());
+	return CELL_OK;
+}
+
+int sys_event_flag_set(u32 eflag_id, u64 bitptn)
+{
+	sys_event_flag.Warning("sys_event_flag_set(eflag_id=0x%x, bitptn=0x%llx)", eflag_id, bitptn);
+
+	event_flag* ef;
+	if(!sys_event_flag.CheckId(eflag_id, ef)) return CELL_ESRCH;
+
+	ef->flags |= bitptn;
+
+	return CELL_OK;
+}
+
+int sys_event_flag_clear(u32 eflag_id, u64 bitptn)
+{
+	sys_event_flag.Warning("sys_event_flag_clear(eflag_id=0x%x, bitptn=0x%llx)", eflag_id, bitptn);
+
+	event_flag* ef;
+	if(!sys_event_flag.CheckId(eflag_id, ef)) return CELL_ESRCH;
+
+	ef->flags &= bitptn;
+
+	return CELL_OK;
+}
+
+int sys_event_flag_cancel(u32 eflag_id, mem32_t num)
+{
+	sys_event_flag.Error("sys_event_flag_cancel(eflag_id=0x%x, num_addr=0x%x)", eflag_id, num.GetAddr());
+	return CELL_OK;
+}
+
+int sys_event_flag_get(u32 eflag_id, mem64_t flags)
+{
+	sys_event_flag.Warning("sys_event_flag_get(eflag_id=0x%x, flags_addr=0x%x)", eflag_id, flags.GetAddr());
+	
+	if (!flags.IsGood())
+	{
+		return CELL_EFAULT;
+	}
+	
+	event_flag* ef;
+	if(!sys_event_flag.CheckId(eflag_id, ef)) return CELL_ESRCH;
+
+	flags = ef->flags;
+
+	return CELL_OK;
+}
--- a/rpcs3/Emu/SysCalls/lv2/SC_Event_flag.h
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Event_flag.h
@ -0,0 +1,31 @@
+#pragma once
+
+enum
+{
+	SYS_SYNC_WAITER_SINGLE = 0x10000,
+	SYS_SYNC_WAITER_MULTIPLE = 0x20000,
+};
+
+struct sys_event_flag_attr
+{
+	be_t<u32> protocol;
+	be_t<u32> pshared;
+	be_t<u64> ipc_key;
+	be_t<int> flags;
+	be_t<int> type;
+	char name[8];
+};
+
+struct event_flag
+{
+	std::atomic<u64> flags;
+	const u32 m_protocol;
+	const int m_type;
+
+	event_flag(u64 pattern, u32 protocol, int type)
+		: flags(pattern)
+		, m_protocol(protocol)
+		, m_type(type)
+	{
+	}
+};
--- a/rpcs3/Emu/SysCalls/lv2/SC_GCM.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_GCM.cpp
@ -7,7 +7,7 @@ extern gcmInfo gcm_info;

 int cellGcmCallback(u32 context_addr, u32 count)
 {
-	GSLockCurrent gslock(GS_LOCK_WAIT_FLUSH);
+	GSLockCurrent gslock(GS_LOCK_WAIT_FLUSH); // could stall on exit

 	CellGcmContextData& ctx = (CellGcmContextData&)Memory[context_addr];
 	CellGcmControl& ctrl = (CellGcmControl&)Memory[gcm_info.control_addr];
--- a/rpcs3/Emu/SysCalls/lv2/SC_Heap.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Heap.cpp
@ -18,9 +18,12 @@ struct HeapInfo
 };

 int sys_heap_create_heap(const u32 heap_addr, const u32 align, const u32 size)
-{
+{	
 	sc_heap.Warning("sys_heap_create_heap(heap_addr=0x%x, align=0x%x, size=0x%x)", heap_addr, align, size);
-	return sc_heap.GetNewId(new HeapInfo(heap_addr, align, size));
+
+	u32 heap_id = sc_heap.GetNewId(new HeapInfo(heap_addr, align, size));
+	sc_heap.Warning("*** sys_heap created(): id=0x%x", heap_id);
+	return heap_id;
 }

 int sys_heap_malloc(const u32 heap_id, const u32 size)
@ -30,5 +33,15 @@ int sys_heap_malloc(const u32 heap_id, const u32 size)
 	HeapInfo* heap;
 	if(!sc_heap.CheckId(heap_id, heap)) return CELL_ESRCH;

-	return Memory.Alloc(size, heap->align);
+	return Memory.Alloc(size, 1);
+}
+
+int _sys_heap_memalign(u32 heap_id, u32 align, u32 size, u64 p4)
+{
+	sc_heap.Warning("_sys_heap_memalign(heap_id=0x%x, align=0x%x, size=0x%x, p4=0x%llx, ... ???)", heap_id, align, size, p4);
+
+	HeapInfo* heap;
+	if(!sc_heap.CheckId(heap_id, heap)) return CELL_ESRCH;
+
+	return Memory.Alloc(size, align);
 }
--- a/rpcs3/Emu/SysCalls/lv2/SC_Lwcond.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Lwcond.cpp
@ -7,25 +7,16 @@ SysCallBase sys_lwcond("sys_lwcond");

 int sys_lwcond_create(mem_ptr_t<sys_lwcond_t> lwcond, mem_ptr_t<sys_lwmutex_t> lwmutex, mem_ptr_t<sys_lwcond_attribute_t> attr)
 {
-	sys_lwcond.Warning("sys_lwcond_create(lwcond_addr=0x%x, lwmutex_addr=0x%x, attr_addr=0x%x)",
+	sys_lwcond.Log("sys_lwcond_create(lwcond_addr=0x%x, lwmutex_addr=0x%x, attr_addr=0x%x)",
 		lwcond.GetAddr(), lwmutex.GetAddr(), attr.GetAddr());

 	if (!lwcond.IsGood() || !lwmutex.IsGood() || !attr.IsGood()) return CELL_EFAULT;

-	u32 protocol = (u32)lwmutex->attribute & SYS_SYNC_ATTR_PROTOCOL_MASK;
-	switch (protocol)
-	{
-	case SYS_SYNC_PRIORITY: break;
-	case SYS_SYNC_RETRY: sys_lwcond.Error("Invalid SYS_SYNC_RETRY attr"); break;
-	case SYS_SYNC_PRIORITY_INHERIT: sys_lwcond.Warning("TODO: SYS_SYNC_PRIORITY_INHERIT attr"); break;
-	case SYS_SYNC_FIFO: break;
-	default: sys_lwcond.Error("Invalid lwmutex protocol(%d)", protocol); break;
-	}
+	lwcond->lwmutex = lwmutex.GetAddr();
+	lwcond->lwcond_queue = sys_lwcond.GetNewId(new LWCond(attr->name_u64));

-	lwcond->lwmutex_addr = lwmutex.GetAddr();
-	lwcond->lwcond_queue = sys_lwcond.GetNewId(new LWCond(protocol, *(u64*)&attr->name));
-
-	sys_lwcond.Warning("*** lwcond created [%s] (protocol=0x%x): id=%d", attr->name, protocol, (u32)lwcond->lwcond_queue);
+	sys_lwcond.Warning("*** lwcond created [%s] (attr=0x%x, lwmutex.sq=0x%x): id=0x%x", 
+		attr->name, (u32)lwmutex->attribute, (u32)lwmutex->sleep_queue, (u32)lwcond->lwcond_queue);
 	return CELL_OK;
 }

@ -44,21 +35,21 @@ int sys_lwcond_destroy(mem_ptr_t<sys_lwcond_t> lwcond)

 int sys_lwcond_signal(mem_ptr_t<sys_lwcond_t> lwcond)
 {
-	sys_lwcond.Warning("sys_lwcond_signal(lwcond_addr=0x%x)", lwcond.GetAddr());
+	sys_lwcond.Log("sys_lwcond_signal(lwcond_addr=0x%x)", lwcond.GetAddr());

 	if (!lwcond.IsGood()) return CELL_EFAULT;
 	LWCond* lwc;
 	u32 id = (u32)lwcond->lwcond_queue;
 	if (!sys_lwcond.CheckId(id, lwc)) return CELL_ESRCH;

-	lwc->signal();
+	lwc->signal(mem_ptr_t<sys_lwmutex_t>(lwcond->lwmutex)->attribute);

 	return CELL_OK;
 }

 int sys_lwcond_signal_all(mem_ptr_t<sys_lwcond_t> lwcond)
 {
-	sys_lwcond.Warning("sys_lwcond_signal_all(lwcond_addr=0x%x)", lwcond.GetAddr());
+	sys_lwcond.Log("sys_lwcond_signal_all(lwcond_addr=0x%x)", lwcond.GetAddr());

 	if (!lwcond.IsGood()) return CELL_EFAULT;
 	LWCond* lwc;
@ -72,7 +63,7 @@ int sys_lwcond_signal_all(mem_ptr_t<sys_lwcond_t> lwcond)

 int sys_lwcond_signal_to(mem_ptr_t<sys_lwcond_t> lwcond, u32 ppu_thread_id)
 {
-	sys_lwcond.Warning("sys_lwcond_signal_to(lwcond_addr=0x%x, ppu_thread_id=%d)", lwcond.GetAddr(), ppu_thread_id);
+	sys_lwcond.Log("sys_lwcond_signal_to(lwcond_addr=0x%x, ppu_thread_id=%d)", lwcond.GetAddr(), ppu_thread_id);

 	if (!lwcond.IsGood()) return CELL_EFAULT;
 	LWCond* lwc;
@ -86,16 +77,17 @@ int sys_lwcond_signal_to(mem_ptr_t<sys_lwcond_t> lwcond, u32 ppu_thread_id)

 int sys_lwcond_wait(mem_ptr_t<sys_lwcond_t> lwcond, u64 timeout)
 {
-	sys_lwcond.Warning("sys_lwcond_wait(lwcond_addr=0x%x, timeout=%llu)", lwcond.GetAddr(), timeout);
+	sys_lwcond.Log("sys_lwcond_wait(lwcond_addr=0x%x, timeout=%llu)", lwcond.GetAddr(), timeout);

 	if (!lwcond.IsGood()) return CELL_EFAULT;
 	LWCond* lwc;
 	u32 id = (u32)lwcond->lwcond_queue;
 	if (!sys_lwcond.CheckId(id, lwc)) return CELL_ESRCH;
 	const u32 tid = GetCurrentPPUThread().GetId();
-	mem_ptr_t<sys_lwmutex_t> lwmutex((u32)lwcond->lwmutex_addr);

-	if ((u32)lwmutex->owner != tid) return CELL_EPERM; // caller must own this lwmutex
+	mem_ptr_t<sys_lwmutex_t> lwmutex(lwcond->lwmutex);
+
+	if ((u32)lwmutex->owner.GetOwner() != tid) return CELL_EPERM; // caller must own this lwmutex
 	lwc->begin_waiting(tid);

 	u32 counter = 0;
--- a/rpcs3/Emu/SysCalls/lv2/SC_Lwcond.h
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Lwcond.h
@ -2,12 +2,16 @@

 struct sys_lwcond_attribute_t
 {
-	char name[8];
+	union
+	{
+		char name[8];
+		u64 name_u64;
+	};
 };

 struct sys_lwcond_t
 {
-	be_t<u32> lwmutex_addr;
+	be_t<u32> lwmutex;
 	be_t<u32> lwcond_queue;
 };

@ -18,22 +22,20 @@ struct LWCond
 	std::mutex m_lock;
 	Array<u32> waiters; // list of waiting threads
 	Array<u32> signaled; // list of signaled threads
-	u32 m_protocol; // protocol
 	u64 m_name; // not used

-	LWCond(u32 prot, u64 name)
+	LWCond(u64 name)
 		: m_name(name)
-		, m_protocol(prot)
 	{
 	}

-	void signal()
+	void signal(u32 _protocol)
 	{
 		std::lock_guard<std::mutex> lock(m_lock);

 		if (waiters.GetCount())
 		{
-			if (m_protocol == SYS_SYNC_PRIORITY)
+			if ((_protocol & SYS_SYNC_ATTR_PROTOCOL_MASK) == SYS_SYNC_PRIORITY)
 			{
 				u64 max_prio = 0;
 				u32 sel = 0;
--- a/rpcs3/Emu/SysCalls/lv2/SC_Lwmutex.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Lwmutex.cpp
@ -5,125 +5,87 @@

 SysCallBase sc_lwmutex("sys_lwmutex");

-std::mutex g_lwmutex;
-
 int sys_lwmutex_create(mem_ptr_t<sys_lwmutex_t> lwmutex, mem_ptr_t<sys_lwmutex_attribute_t> attr)
 {
-	sc_lwmutex.Warning("sys_lwmutex_create(lwmutex_addr=0x%x, lwmutex_attr_addr=0x%x)", 
+	sc_lwmutex.Log("sys_lwmutex_create(lwmutex_addr=0x%x, lwmutex_attr_addr=0x%x)", 
 		lwmutex.GetAddr(), attr.GetAddr());

 	if (!lwmutex.IsGood() || !attr.IsGood()) return CELL_EFAULT;

-	switch ((u32)attr->attr_recursive)
+	switch (attr->attr_recursive.ToBE())
 	{
-	case SYS_SYNC_RECURSIVE: break;
-	case SYS_SYNC_NOT_RECURSIVE: break;
-	default: return CELL_EINVAL;
+	case se32(SYS_SYNC_RECURSIVE): break;
+	case se32(SYS_SYNC_NOT_RECURSIVE): break;
+	default: sc_lwmutex.Error("Unknown 0x%x recursive attr", (u32)attr->attr_recursive); return CELL_EINVAL;
 	}

-	switch ((u32)attr->attr_protocol)
+	switch (attr->attr_protocol.ToBE())
 	{
-	case SYS_SYNC_PRIORITY: sc_lwmutex.Warning("TODO: SYS_SYNC_PRIORITY attr"); break;
-	case SYS_SYNC_RETRY: sc_lwmutex.Warning("TODO: SYS_SYNC_RETRY attr"); break;
-	case SYS_SYNC_PRIORITY_INHERIT: sc_lwmutex.Warning("TODO: SYS_SYNC_PRIORITY_INHERIT attr"); break;
-	case SYS_SYNC_FIFO: sc_lwmutex.Warning("TODO: SYS_SYNC_FIFO attr"); break;
-	default: return CELL_EINVAL;
+	case se32(SYS_SYNC_PRIORITY): break;
+	case se32(SYS_SYNC_RETRY): break;
+	case se32(SYS_SYNC_PRIORITY_INHERIT): sc_lwmutex.Error("Invalid SYS_SYNC_PRIORITY_INHERIT protocol attr"); return CELL_EINVAL;
+	case se32(SYS_SYNC_FIFO): break;
+	default: sc_lwmutex.Error("Unknown 0x%x protocol attr", (u32)attr->attr_protocol); return CELL_EINVAL;
 	}

-	lwmutex->attribute = (u32)attr->attr_protocol | (u32)attr->attr_recursive;
+	lwmutex->attribute = attr->attr_protocol | attr->attr_recursive;
 	lwmutex->all_info = 0;
 	lwmutex->pad = 0;
 	lwmutex->recursive_count = 0;
-	lwmutex->sleep_queue = 0;

-	sc_lwmutex.Warning("*** lwmutex created [%s] (attribute=0x%x): id=???", attr->name, (u32)lwmutex->attribute);
+	u32 sq_id = sc_lwmutex.GetNewId(new SleepQueue(attr->name_u64));
+	lwmutex->sleep_queue = sq_id;
+
+	sc_lwmutex.Log("*** lwmutex created [%s] (attribute=0x%x): sleep_queue=0x%x", 
+		attr->name, (u32)lwmutex->attribute, sq_id);

 	return CELL_OK;
 }

 int sys_lwmutex_destroy(mem_ptr_t<sys_lwmutex_t> lwmutex)
 {
-	sc_lwmutex.Warning("sys_lwmutex_destroy(lwmutex_addr=0x%x)", lwmutex.GetAddr());
+	sc_lwmutex.Log("sys_lwmutex_destroy(lwmutex_addr=0x%x)", lwmutex.GetAddr());

 	if (!lwmutex.IsGood()) return CELL_EFAULT;

-	if (!lwmutex->attribute) return CELL_EINVAL;
+	u32 sq_id = lwmutex->sleep_queue;
+	if (!Emu.GetIdManager().CheckID(sq_id)) return CELL_ESRCH;

-	{ // global lock
-		std::lock_guard<std::mutex> lock(g_lwmutex);
-
-		if (!lwmutex->owner)
-		{
-			lwmutex->owner = ~0; // make it unable to lock
-			lwmutex->attribute = 0;
-		}
-		else
-		{
-			return CELL_EBUSY;
-		}
+	// try to make it unable to lock
+	switch (int res = lwmutex->trylock(~0)) 
+	{
+	case CELL_OK:
+		lwmutex->attribute = 0;
+		lwmutex->sleep_queue = 0;
+		Emu.GetIdManager().RemoveID(sq_id);
+	default: return res;
 	}
-
-	return CELL_OK;
 }

 int sys_lwmutex_lock(mem_ptr_t<sys_lwmutex_t> lwmutex, u64 timeout)
 {
-	sc_lwmutex.Warning("sys_lwmutex_lock(lwmutex_addr=0x%x, timeout=%lld)", lwmutex.GetAddr(), timeout);
+	sc_lwmutex.Log("sys_lwmutex_lock(lwmutex_addr=0x%x, timeout=%lld)", lwmutex.GetAddr(), timeout);

 	if (!lwmutex.IsGood()) return CELL_EFAULT;

-	if (!lwmutex->attribute) return CELL_EINVAL;
-
-	const u32 tid = GetCurrentPPUThread().GetId();
-
-	int res = lwmutex->trylock(tid);
-	if (res != CELL_EBUSY) return res;
-
-	u32 counter = 0;
-	const u32 max_counter = timeout ? (timeout / 1000) : 20000;
-	do // waiting
-	{
-		if (Emu.IsStopped()) return CELL_ETIMEDOUT;
-		Sleep(1);
-
-		res = lwmutex->trylock(tid);
-		if (res != CELL_EBUSY) return res;
-		if (!lwmutex->attribute) return CELL_EINVAL;
-
-		if (counter++ > max_counter)
-		{
-			if (!timeout) 
-			{
-				sc_lwmutex.Warning("sys_lwmutex_lock(lwmutex_addr=0x%x): TIMEOUT", lwmutex.GetAddr());
-				counter = 0;
-			}
-			else
-			{
-				return CELL_ETIMEDOUT;
-			}
-		}
-	} while (true);
+	return lwmutex->lock(GetCurrentPPUThread().GetId(), timeout ? ((timeout < 1000) ? 1 : (timeout / 1000)) : 0);
 }

 int sys_lwmutex_trylock(mem_ptr_t<sys_lwmutex_t> lwmutex)
 {
-	sc_lwmutex.Warning("sys_lwmutex_trylock(lwmutex_addr=0x%x)", lwmutex.GetAddr());
+	sc_lwmutex.Log("sys_lwmutex_trylock(lwmutex_addr=0x%x)", lwmutex.GetAddr());

 	if (!lwmutex.IsGood()) return CELL_EFAULT;

-	if (!lwmutex->attribute) return CELL_EINVAL;
-
 	return lwmutex->trylock(GetCurrentPPUThread().GetId());
 }

 int sys_lwmutex_unlock(mem_ptr_t<sys_lwmutex_t> lwmutex)
 {
-	sc_lwmutex.Warning("sys_lwmutex_unlock(lwmutex_addr=0x%x)", lwmutex.GetAddr());
+	sc_lwmutex.Log("sys_lwmutex_unlock(lwmutex_addr=0x%x)", lwmutex.GetAddr());

 	if (!lwmutex.IsGood()) return CELL_EFAULT;

-	if (!lwmutex->unlock(GetCurrentPPUThread().GetId())) return CELL_EPERM;
-
-	return CELL_OK;
+	return lwmutex->unlock(GetCurrentPPUThread().GetId());
 }

--- a/rpcs3/Emu/SysCalls/lv2/SC_Lwmutex.h
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Lwmutex.h
@ -1,4 +1,5 @@
 #pragma once
+#include <Utilities/SMutex.h>

 // attr_protocol (waiting scheduling policy)
 enum
@ -9,7 +10,7 @@ enum
 	SYS_SYNC_PRIORITY = 2,
 	// Basic Priority Inheritance Protocol
 	SYS_SYNC_PRIORITY_INHERIT = 3,
-	// ????
+	// Not selected while unlocking
 	SYS_SYNC_RETRY = 4,
 	//
 	SYS_SYNC_ATTR_PROTOCOL_MASK = 0xF, 
@ -30,10 +31,104 @@ struct sys_lwmutex_attribute_t
 {
 	be_t<u32> attr_protocol;
 	be_t<u32> attr_recursive;
-	char name[8];
+	union
+	{
+		char name[8];
+		u64 name_u64;
+	};
 };

-extern std::mutex g_lwmutex;
+class SleepQueue
+{
+	/* struct q_rec
+	{
+		u32 tid;
+		u64 prio;
+		q_rec(u32 tid, u64 prio): tid(tid), prio(prio) {}
+	}; */
+
+	SMutex m_mutex;
+	Array<u32> list;
+	u64 m_name;
+
+public:
+	SleepQueue(u64 name)
+		: m_name(name)
+	{
+	}
+
+	void push(u32 tid)
+	{
+		SMutexLocker lock(m_mutex);
+		list.AddCpy(tid);
+	}
+
+	u32 pop() // SYS_SYNC_FIFO
+	{
+		SMutexLocker lock(m_mutex);
+		
+		while (true)
+		{
+			if (list.GetCount())
+			{
+				u32 res = list[0];
+				list.RemoveAt(0);
+				if (Emu.GetIdManager().CheckID(res))
+				// check thread
+				{
+					return res;
+				}
+			}
+			return 0;
+		};
+	}
+
+	u32 pop_prio() // SYS_SYNC_PRIORITY
+	{
+		SMutexLocker lock(m_mutex);
+
+		while (true)
+		{
+			if (list.GetCount())
+			{
+				u64 max_prio = 0;
+				u32 sel = 0;
+				for (u32 i = 0; i < list.GetCount(); i++)
+				{
+					CPUThread* t = Emu.GetCPU().GetThread(list[i]);
+					if (!t)
+					{
+						list[i] = 0;
+						sel = i;
+						break;
+					}
+
+					u64 prio = t->GetPrio();
+					if (prio > max_prio)
+					{
+						max_prio = prio;
+						sel = i;
+					}
+				}
+				u32 res = list[sel];
+				list.RemoveAt(sel);
+				/* if (Emu.GetIdManager().CheckID(res)) */
+				if (res)
+				// check thread
+				{
+					return res;
+				}
+			}
+			return 0;
+		}
+	}
+
+	u32 pop_prio_inherit() // (TODO)
+	{
+		ConLog.Error("TODO: SleepQueue::pop_prio_inherit()");
+		Emu.Pause();
+	}
+};

 struct sys_lwmutex_t
 {
@ -41,8 +136,8 @@ struct sys_lwmutex_t
 	{
 		struct // sys_lwmutex_lock_info_t
 		{
-			/* volatile */ be_t<u32> owner;
-			/* volatile */ be_t<u32> waiter;
+			/* volatile */ SMutexBE owner;
+			/* volatile */ be_t<u32> waiter; // not used
 		};
 		struct
 		{ 
@ -54,77 +149,103 @@ struct sys_lwmutex_t
 	be_t<u32> sleep_queue;
 	be_t<u32> pad;

-	int trylock(u32 tid)
+	int trylock(be_t<u32> tid)
 	{
-		std::lock_guard<std::mutex> lock(g_lwmutex); // global lock
+		if (!attribute.ToBE()) return CELL_EINVAL;

-		if ((u32)attribute & SYS_SYNC_RECURSIVE)
+		if (tid == owner.GetOwner()) 
 		{
-			if (tid == (u32)owner)
+			if (attribute.ToBE() & se32(SYS_SYNC_RECURSIVE))
 			{
-				recursive_count = (u32)recursive_count + 1;
-				if ((u32)recursive_count == 0xffffffff) return CELL_EKRESOURCE;
+				recursive_count += 1;
+				if (!recursive_count.ToBE()) return CELL_EKRESOURCE;
 				return CELL_OK;
 			}
-		}
-		else // recursive not allowed
-		{
-			if (tid == (u32)owner)
+			else
 			{
 				return CELL_EDEADLK;
 			}
 		}

-		if (!(u32)owner) // try lock
+		switch (owner.trylock(tid))
 		{
-			owner = tid; 
-			recursive_count = 1;
-			return CELL_OK;
-		}
-		else
-		{
-			return CELL_EBUSY;
+		case SMR_OK: recursive_count = 1; return CELL_OK;
+		case SMR_FAILED: return CELL_EBUSY;
+		default: return CELL_EINVAL;
 		}
 	}

-	bool unlock(u32 tid)
+	int unlock(be_t<u32> tid)
 	{
-		std::lock_guard<std::mutex> lock(g_lwmutex); // global lock
-
-		if (tid != (u32)owner)
+		if (tid != owner.GetOwner())
 		{
-			return false;
+			return CELL_EPERM;
 		}
 		else
 		{
-			recursive_count = (u32)recursive_count - 1;
-			if (!(u32)recursive_count)
+			recursive_count -= 1;
+			if (!recursive_count.ToBE())
 			{
-				waiter = 0; // not used yet
-				owner = 0; // release
+				be_t<u32> target = 0;
+				switch (attribute.ToBE() & se32(SYS_SYNC_ATTR_PROTOCOL_MASK))
+				{
+				case se32(SYS_SYNC_FIFO):
+				case se32(SYS_SYNC_PRIORITY):
+					SleepQueue* sq;
+					if (!Emu.GetIdManager().GetIDData(sleep_queue, sq)) return CELL_ESRCH;
+					target = attribute.ToBE() & se32(SYS_SYNC_FIFO) ? sq->pop() : sq->pop_prio();
+				case se32(SYS_SYNC_RETRY): default: owner.unlock(tid, target); break;
+				}
 			}
-			return true;
+			return CELL_OK;
+		}
+	}
+
+	int lock(be_t<u32> tid, u64 timeout)
+	{
+		switch (int res = trylock(tid))
+		{
+		case CELL_EBUSY: break;
+		default: return res;
+		}
+
+		switch (attribute.ToBE() & se32(SYS_SYNC_ATTR_PROTOCOL_MASK))
+		{
+		case se32(SYS_SYNC_PRIORITY):
+		case se32(SYS_SYNC_FIFO):
+			SleepQueue* sq;
+			if (!Emu.GetIdManager().GetIDData(sleep_queue, sq)) return CELL_ESRCH;
+			sq->push(tid);
+		default: break;
+		}
+
+		switch (owner.lock(tid, timeout))
+		{
+		case SMR_OK: case SMR_SIGNAL: recursive_count = 1; return CELL_OK;
+		case SMR_TIMEOUT: return CELL_ETIMEDOUT;
+		default: return CELL_EINVAL;
 		}
 	}
 };

-struct lwmutex_locker
+class lwmutex_locker
 {
-private:
 	mem_ptr_t<sys_lwmutex_t> m_mutex;
-	u32 m_id;
-public:
-	const int res;	
+	be_t<u32> m_id;

-	lwmutex_locker(u32 lwmutex_addr, u32 tid)
+	lwmutex_locker(mem_ptr_t<sys_lwmutex_t> lwmutex, be_t<u32> tid, u64 timeout = 0)
 		: m_id(tid)
-		, m_mutex(lwmutex_addr)
-		, res(m_mutex->trylock(m_id))
+		, m_mutex(lwmutex)
 	{
+		if (int res = m_mutex->lock(m_id, timeout))
+		{
+			ConLog.Error("lwmutex_locker: m_mutex->lock failed(res=0x%x)", res);
+			Emu.Pause();
+		}
 	}

 	~lwmutex_locker()
 	{
-		if (res == CELL_OK) m_mutex->unlock(m_id);
+		m_mutex->unlock(m_id);
 	}
 };
--- a/rpcs3/Emu/SysCalls/lv2/SC_Memory.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Memory.cpp
@ -6,7 +6,7 @@ SysCallBase sc_mem("memory");

 int sys_memory_container_create(u32 cid_addr, u32 yield_size)
 {
-	sc_mem.Warning("(HACK!) sys_memory_container_create(cid_addr=0x%x,yield_size=0x%x)", cid_addr, yield_size);
+	sc_mem.Warning("sys_memory_container_create(cid_addr=0x%x,yield_size=0x%x)", cid_addr, yield_size);

 	if(!Memory.IsGoodAddr(cid_addr, 4))
 	{
@ -15,17 +15,13 @@ int sys_memory_container_create(u32 cid_addr, u32 yield_size)

 	yield_size &= ~0xfffff; //round down to 1 MB granularity

-	//alignment hack (Memory.Alloc does not support alignment yet): alloc size is increased
-	u64 addr = Memory.Alloc(yield_size + 0x100000, 0x100000); //1 MB alignment (???)
+	u64 addr = Memory.Alloc(yield_size, 0x100000); //1 MB alignment

 	if(!addr)
 	{
 		return CELL_ENOMEM;
 	}

-	//fix alignment:
-	addr = (addr + 0x100000) & ~0xfffff;
-
 	Memory.Write32(cid_addr, sc_mem.GetNewId(new MemoryContainerInfo(addr, yield_size)));
 	return CELL_OK;
 }
@ -49,20 +45,18 @@ int sys_memory_container_destroy(u32 cid)
 int sys_memory_allocate(u32 size, u32 flags, u32 alloc_addr_addr)
 {
 	//0x30000100;
-	sc_mem.Warning("(HACK!) sys_memory_allocate(size=0x%x, flags=0x%x)", size, flags);
+	sc_mem.Log("sys_memory_allocate(size=0x%x, flags=0x%x)", size, flags);
 	u32 addr;
 	switch(flags)
 	{
 	case SYS_MEMORY_PAGE_SIZE_1M:
 		if(size & 0xfffff) return CELL_EALIGN;
-		addr = Memory.Alloc(size + 0x100000, 0x100000);
-		addr = (addr + 0x100000) & ~0xfffff;
+		addr = Memory.Alloc(size, 1);
 	break;

 	case SYS_MEMORY_PAGE_SIZE_64K:
 		if(size & 0xffff) return CELL_EALIGN;
-		addr = Memory.Alloc(size + 0x10000, 0x10000);
-		addr = (addr + 0x10000) & ~0xffff;
+		addr = Memory.Alloc(size, 1);
 	break;

 	default: return CELL_EINVAL;
--- a/rpcs3/Emu/SysCalls/lv2/SC_Mutex.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Mutex.cpp
@ -1,12 +1,13 @@
 #include "stdafx.h"
 #include "Emu/SysCalls/SysCalls.h"
 #include "SC_Mutex.h"
+#include "Utilities/SMutex.h"

 SysCallBase sys_mtx("sys_mutex");

 int sys_mutex_create(u32 mutex_id_addr, u32 attr_addr)
 {
-	sys_mtx.Log("sys_mutex_create(mutex_id_addr=0x%x, attr_addr=0x%x)",
+	sys_mtx.Warning("sys_mutex_create(mutex_id_addr=0x%x, attr_addr=0x%x)",
 		mutex_id_addr, attr_addr);

 	if(!Memory.IsGoodAddr(mutex_id_addr) || !Memory.IsGoodAddr(attr_addr)) return CELL_EFAULT;
@ -48,9 +49,28 @@ int sys_mutex_lock(u32 mutex_id, u64 timeout)
 	mutex* mtx_data = nullptr;
 	if(!sys_mtx.CheckId(mutex_id, mtx_data)) return CELL_ESRCH;

-	mtx_data->mtx.Lock();
+	u32 counter = 0;
+	const u32 max_counter = timeout ? (timeout / 1000) : 20000;
+	do
+	{
+		if (Emu.IsStopped()) return CELL_ETIMEDOUT;

-	return CELL_OK;
+		if (mtx_data->mtx.TryLock() == wxMUTEX_NO_ERROR) return CELL_OK;
+		Sleep(1);
+
+		if (counter++ > max_counter)
+		{
+			if (!timeout) 
+			{
+				sys_mtx.Warning("sys_mutex_lock(mutex_id=0x%x, timeout=0x%llx): TIMEOUT", mutex_id, timeout);
+				counter = 0;
+			}
+			else
+			{
+				return CELL_ETIMEDOUT;
+			}
+		}		
+	} while (true);
 }

 int sys_mutex_trylock(u32 mutex_id)
@ -60,7 +80,7 @@ int sys_mutex_trylock(u32 mutex_id)
 	mutex* mtx_data = nullptr;
 	if(!sys_mtx.CheckId(mutex_id, mtx_data)) return CELL_ESRCH;

-	if(mtx_data->mtx.TryLock()) return 1;
+	if (mtx_data->mtx.TryLock() != wxMUTEX_NO_ERROR) return CELL_EBUSY;

 	return CELL_OK;
 }
--- a/rpcs3/Emu/SysCalls/lv2/SC_PPU_Thread.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_PPU_Thread.cpp
@ -131,7 +131,7 @@ int sys_ppu_thread_restart(u32 thread_id)
 	return CELL_OK;
 }

-int sys_ppu_thread_create(u32 thread_id_addr, u32 entry, u32 arg, int prio, u32 stacksize, u64 flags, u32 threadname_addr)
+int sys_ppu_thread_create(u32 thread_id_addr, u32 entry, u64 arg, int prio, u32 stacksize, u64 flags, u32 threadname_addr)
 {
 	sysPrxForUser.Log("sys_ppu_thread_create(thread_id_addr=0x%x, entry=0x%x, arg=0x%x, prio=%d, stacksize=0x%x, flags=0x%llx, threadname_addr=0x%x('%s'))",
 		thread_id_addr, entry, arg, prio, stacksize, flags, threadname_addr, Memory.ReadString(threadname_addr).mb_str());
--- a/rpcs3/Emu/SysCalls/lv2/SC_Process.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Process.cpp
@ -40,11 +40,8 @@ int sys_process_getppid()
 int sys_process_exit(int errorcode)
 {
 	sc_p.Warning("sys_process_exit(%d)", errorcode);
-#ifdef _DEBUG
-	Emu.Pause();
-#else
-	Emu.Stop();
-#endif
+	Emu.Pause(); // Emu.Stop() does crash
+	ConLog.Success("Process finished");
 	return CELL_OK;
 }

--- a/rpcs3/Emu/SysCalls/lv2/SC_Rwlock.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Rwlock.cpp
@ -10,23 +10,22 @@ int sys_rwlock_create(mem32_t rw_lock_id, mem_ptr_t<sys_rwlock_attribute_t> attr

 	if (!rw_lock_id.IsGood() || !attr.IsGood()) return CELL_EFAULT;

-	switch ((u32)attr->attr_protocol)
+	switch (attr->attr_protocol.ToBE())
 	{
-	case SYS_SYNC_PRIORITY: sys_rwlock.Warning("TODO: SYS_SYNC_PRIORITY attr"); break;
-	case SYS_SYNC_RETRY: sys_rwlock.Error("Invalid SYS_SYNC_RETRY attr"); break;
-	case SYS_SYNC_PRIORITY_INHERIT: sys_rwlock.Warning("TODO: SYS_SYNC_PRIORITY_INHERIT attr"); break;
-	case SYS_SYNC_FIFO: break;
+	case se32(SYS_SYNC_PRIORITY): sys_rwlock.Warning("TODO: SYS_SYNC_PRIORITY attr"); break;
+	case se32(SYS_SYNC_RETRY): sys_rwlock.Error("Invalid SYS_SYNC_RETRY attr"); break;
+	case se32(SYS_SYNC_PRIORITY_INHERIT): sys_rwlock.Warning("TODO: SYS_SYNC_PRIORITY_INHERIT attr"); break;
+	case se32(SYS_SYNC_FIFO): break;
 	default: return CELL_EINVAL;
 	}

-	if ((u32)attr->attr_pshared != 0x200)
+	if (attr->attr_pshared.ToBE() != se32(0x200))
 	{
 		sys_rwlock.Error("Invalid attr_pshared(0x%x)", (u32)attr->attr_pshared);
 		return CELL_EINVAL;
 	}

-	rw_lock_id = sys_rwlock.GetNewId(new RWLock((u32)attr->attr_protocol, (u32)attr->attr_pshared, 
-		(u64)attr->key, (s32)attr->flags, *(u64*)&attr->name));
+	rw_lock_id = sys_rwlock.GetNewId(new RWLock((u32)attr->attr_protocol, attr->name_u64));

 	sys_rwlock.Warning("*** rwlock created [%s] (protocol=0x%x): id=%d", 
 		attr->name, (u32)attr->attr_protocol, rw_lock_id.GetValue());
@ -52,7 +51,7 @@ int sys_rwlock_destroy(u32 rw_lock_id)

 int sys_rwlock_rlock(u32 rw_lock_id, u64 timeout)
 {
-	sys_rwlock.Warning("sys_rwlock_rlock(rw_lock_id=%d, timeout=%llu)", rw_lock_id, timeout);
+	sys_rwlock.Log("sys_rwlock_rlock(rw_lock_id=%d, timeout=%llu)", rw_lock_id, timeout);

 	RWLock* rw;
 	if (!sys_rwlock.CheckId(rw_lock_id, rw)) return CELL_ESRCH;
@ -86,7 +85,7 @@ int sys_rwlock_rlock(u32 rw_lock_id, u64 timeout)

 int sys_rwlock_tryrlock(u32 rw_lock_id)
 {
-	sys_rwlock.Warning("sys_rwlock_tryrlock(rw_lock_id=%d)", rw_lock_id);
+	sys_rwlock.Log("sys_rwlock_tryrlock(rw_lock_id=%d)", rw_lock_id);

 	RWLock* rw;
 	if (!sys_rwlock.CheckId(rw_lock_id, rw)) return CELL_ESRCH;
@ -98,7 +97,7 @@ int sys_rwlock_tryrlock(u32 rw_lock_id)

 int sys_rwlock_runlock(u32 rw_lock_id)
 {
-	sys_rwlock.Warning("sys_rwlock_runlock(rw_lock_id=%d)", rw_lock_id);
+	sys_rwlock.Log("sys_rwlock_runlock(rw_lock_id=%d)", rw_lock_id);

 	RWLock* rw;
 	if (!sys_rwlock.CheckId(rw_lock_id, rw)) return CELL_ESRCH;
@ -110,7 +109,7 @@ int sys_rwlock_runlock(u32 rw_lock_id)

 int sys_rwlock_wlock(u32 rw_lock_id, u64 timeout)
 {
-	sys_rwlock.Warning("sys_rwlock_wlock(rw_lock_id=%d, timeout=%llu)", rw_lock_id, timeout);
+	sys_rwlock.Log("sys_rwlock_wlock(rw_lock_id=%d, timeout=%llu)", rw_lock_id, timeout);

 	RWLock* rw;
 	if (!sys_rwlock.CheckId(rw_lock_id, rw)) return CELL_ESRCH;
@ -146,7 +145,7 @@ int sys_rwlock_wlock(u32 rw_lock_id, u64 timeout)

 int sys_rwlock_trywlock(u32 rw_lock_id)
 {
-	sys_rwlock.Warning("sys_rwlock_trywlock(rw_lock_id=%d)", rw_lock_id);
+	sys_rwlock.Log("sys_rwlock_trywlock(rw_lock_id=%d)", rw_lock_id);

 	RWLock* rw;
 	if (!sys_rwlock.CheckId(rw_lock_id, rw)) return CELL_ESRCH;
@ -161,7 +160,7 @@ int sys_rwlock_trywlock(u32 rw_lock_id)

 int sys_rwlock_wunlock(u32 rw_lock_id)
 {
-	sys_rwlock.Warning("sys_rwlock_wunlock(rw_lock_id=%d)", rw_lock_id);
+	sys_rwlock.Log("sys_rwlock_wunlock(rw_lock_id=%d)", rw_lock_id);

 	RWLock* rw;
 	if (!sys_rwlock.CheckId(rw_lock_id, rw)) return CELL_ESRCH;
--- a/rpcs3/Emu/SysCalls/lv2/SC_Rwlock.h
+++ b/rpcs3/Emu/SysCalls/lv2/SC_Rwlock.h
@ -6,8 +6,12 @@ struct sys_rwlock_attribute_t
 	be_t<u32> attr_pshared; // == 0x200 (NOT SHARED)
 	be_t<u64> key; // process-shared key (not used)
 	be_t<s32> flags; // process-shared flags (not used)
-	be_t<u32> pad;
-	char name[8];
+	be_t<u32> pad; // not used
+	union
+	{
+		char name[8];
+		u64 name_u64;
+	};
 };

 #pragma pack()
@ -20,21 +24,15 @@ struct RWLock
 	Array<u32> rlock_list; // read lock list
 	u32 m_protocol; // TODO

-	u32 m_pshared; // not used
-	u64 m_key; // not used
-	s32 m_flags; // not used
 	union
 	{
-		u64 m_name_data; // not used
+		u64 m_name_u64;
 		char m_name[8];
 	};

-	RWLock(u32 protocol, u32 pshared, u64 key, s32 flags, u64 name)
+	RWLock(u32 protocol, u64 name)
 		: m_protocol(protocol)
-		, m_pshared(pshared)
-		, m_key(key)
-		, m_flags(flags)
-		, m_name_data(name)
+		, m_name_u64(name)
 		, wlock_thread(0)
 	{
 	}
--- a/rpcs3/Emu/SysCalls/lv2/SC_SPU_Thread.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_SPU_Thread.cpp
@ -40,7 +40,7 @@ u32 LoadSpuImage(vfsStream& stream, u32& spu_ep)
 	ELFLoader l(stream);
 	l.LoadInfo();
 	const u32 alloc_size = 256 * 1024 /*0x1000000 - stream.GetSize()*/;
-	u32 spu_offset = Memory.MainMem.Alloc(alloc_size);
+	u32 spu_offset = Memory.MainMem.AllocAlign(alloc_size);
 	l.LoadData(spu_offset);
 	spu_ep = l.GetEntry();
 	return spu_offset;
@ -104,7 +104,7 @@ int sys_spu_thread_initialize(mem32_t thread, u32 group, u32 spu_num, mem_ptr_t<
 		return CELL_ESRCH;
 	}

-	if(!thread.IsGood() || !img.IsGood() || !attr.IsGood() || !attr.IsGood())
+	if(!thread.IsGood() || !img.IsGood() || !attr.IsGood() || !arg.IsGood())
 	{
 		return CELL_EFAULT;
 	}
@ -133,7 +133,7 @@ int sys_spu_thread_initialize(mem32_t thread, u32 group, u32 spu_num, mem_ptr_t<

 	CPUThread& new_thread = Emu.GetCPU().AddThread(CPU_THREAD_SPU);
 	//copy SPU image:
-	u32 spu_offset = Memory.MainMem.Alloc(256 * 1024);
+	u32 spu_offset = Memory.MainMem.AllocAlign(256 * 1024);
 	memcpy(Memory + spu_offset, Memory + (u32)img->segs_addr, 256 * 1024);
 	//initialize from new place:
 	new_thread.SetOffset(spu_offset);
@ -177,24 +177,70 @@ int sys_spu_thread_set_argument(u32 id, mem_ptr_t<sys_spu_thread_argument> arg)
 	return CELL_OK;
 }

-//173
-int sys_spu_thread_group_start(u32 id)
+//165
+int sys_spu_thread_get_exit_status(u32 id, mem32_t status)
 {
-	sc_spu.Warning("sys_spu_thread_group_start(id=0x%x)", id);
+	sc_spu.Warning("sys_spu_thread_get_exit_status(id=0x%x, status_addr=0x%x)", id, status.GetAddr());

-	if(!Emu.GetIdManager().CheckID(id))
+	if (!status.IsGood())
+	{
+		return CELL_EFAULT;
+	}
+
+	CPUThread* thr = Emu.GetCPU().GetThread(id);
+
+	if(!thr || (thr->GetType() != CPU_THREAD_SPU && thr->GetType() != CPU_THREAD_RAW_SPU))
 	{
 		return CELL_ESRCH;
 	}

+	u32 res;
+	if (!(*(SPUThread*)thr).SPU.Out_MBox.Pop(res) || !thr->IsStopped())
+	{
+		return CELL_ESTAT;
+	}
+
+	status = res;
+	return CELL_OK;
+}
+
+//171
+int sys_spu_thread_group_destroy(u32 id)
+{
+	sc_spu.Warning("sys_spu_thread_group_destroy(id=0x%x)", id);
+
 	SpuGroupInfo* group_info;
 	if(!Emu.GetIdManager().GetIDData(id, group_info))
 	{
 		return CELL_ESRCH;
 	}

-	//Emu.Pause();
-	for (int i = 0; i < group_info->list.GetCount(); i++)
+	if (group_info->lock) // ???
+	{
+		return CELL_EBUSY;
+	}
+
+	for (u32 i = 0; i < group_info->list.GetCount(); i++)
+	{
+		Emu.GetCPU().RemoveThread(group_info->list[i]);
+	}
+
+	Emu.GetIdManager().RemoveID(id);
+	return CELL_OK;
+}
+
+//173
+int sys_spu_thread_group_start(u32 id)
+{
+	sc_spu.Warning("sys_spu_thread_group_start(id=0x%x)", id);
+
+	SpuGroupInfo* group_info;
+	if(!Emu.GetIdManager().GetIDData(id, group_info))
+	{
+		return CELL_ESRCH;
+	}
+
+	for (u32 i = 0; i < group_info->list.GetCount(); i++)
 	{
 		CPUThread* t;
 		if (t = Emu.GetCPU().GetThread(group_info->list[i]))
@ -218,10 +264,9 @@ int sys_spu_thread_group_suspend(u32 id)
 	}

 	//Emu.Pause();
-	for (int i = 0; i < group_info->list.GetCount(); i++)
+	for (u32 i = 0; i < group_info->list.GetCount(); i++)
 	{
-		CPUThread* t;
-		if (t = Emu.GetCPU().GetThread(group_info->list[i]))
+		if (CPUThread* t = Emu.GetCPU().GetThread(group_info->list[i]))
 		{
 			t->Pause();
 		}
@ -273,14 +318,17 @@ int sys_spu_thread_group_join(u32 id, mem32_t cause, mem32_t status)
 	cause = SYS_SPU_THREAD_GROUP_JOIN_ALL_THREADS_EXIT;
 	status = 0; //unspecified because of ALL_THREADS_EXIT

-	for (int i = 0; i < group_info->list.GetCount(); i++)
+	for (u32 i = 0; i < group_info->list.GetCount(); i++)
 	{
-		while (Emu.GetCPU().GetThread(group_info->list[i]))
+		while (CPUThread* t = Emu.GetCPU().GetThread(group_info->list[i]))
 		{
-			Sleep(1);
+			if (!t->IsRunning())
+			{
+				break;
+			}
 			if (Emu.IsStopped()) return CELL_OK;
+			Sleep(1);
 		}
-		group_info->list[i] = 0;
 	}

 	group_info->lock = 0; // release lock
@ -514,7 +562,7 @@ int sys_spu_thread_group_connect_event_all_threads(u32 id, u32 eq, u64 req, u32
 		return CELL_ESRCH;
 	}
 	
-	for(int i=0; i<group->list.GetCount(); ++i)
+	for(u32 i=0; i<group->list.GetCount(); ++i)
 	{
 		CPUThread* t;
 		if(t = Emu.GetCPU().GetThread(group->list[i]))
@ -537,3 +585,10 @@ int sys_spu_thread_group_connect_event_all_threads(u32 id, u32 eq, u64 req, u32
 	}
 	return CELL_OK;
 }
+
+int sys_spu_thread_bind_queue(u32 id, u32 spuq, u32 spuq_num)
+{
+	sc_spu.Error("sys_spu_thread_bind_queue(id=0x%x, spuq=0x%x, spuq_num=0x%x)", id, spuq, spuq_num);
+
+	return CELL_OK;
+}
--- a/rpcs3/Emu/SysCalls/lv2/SC_TTY.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_TTY.cpp
@ -11,7 +11,7 @@ int sys_tty_read(u32 ch, u64 buf_addr, u32 len, u64 preadlen_addr)

 int sys_tty_write(u32 ch, u64 buf_addr, u32 len, u64 pwritelen_addr)
 {
-	if(ch < 0 || ch > 15 || len <= 0) return CELL_EINVAL;
+	if(ch < 0 || ch > 15 || (s32)len <= 0) return CELL_EINVAL;
 	if(!Memory.IsGoodAddr(buf_addr)) return CELL_EFAULT;
 	
 	Emu.GetDbgCon().Write(ch, Memory.ReadString(buf_addr, len));
--- a/rpcs3/Emu/SysCalls/lv2/SC_VM.cpp
+++ b/rpcs3/Emu/SysCalls/lv2/SC_VM.cpp
@ -108,7 +108,7 @@ int sys_vm_return_memory(u32 addr, u32 size)
 	}

 	// The memory size to return should not be superior to the virtual size in use minus 1MB.
-	if((current_ct->size - size - 0x100000) < 0)
+	if(current_ct->size < (size + 0x100000))
 	{
 		return CELL_EBUSY;
 	}
@ -130,7 +130,7 @@ int sys_vm_lock(u32 addr, u32 size)
 	}

 	// The memory size to return should not be superior to the virtual size to lock minus 1MB.
-	if((current_ct->size - size - 0x100000) < 0)
+	if(current_ct->size < (size + 0x100000))
 	{
 		return CELL_EBUSY;
 	}
--- a/rpcs3/Emu/System.cpp
+++ b/rpcs3/Emu/System.cpp
@ -305,7 +305,7 @@ void Emulator::Load()
 		ConLog.Write("offset = 0x%llx", Memory.MainMem.GetStartAddr());
 		ConLog.Write("max addr = 0x%x", l.GetMaxAddr());
 		thread.SetOffset(Memory.MainMem.GetStartAddr());
-		Memory.MainMem.Alloc(Memory.MainMem.GetStartAddr() + l.GetMaxAddr(), 0xFFFFED - l.GetMaxAddr());
+		Memory.MainMem.AllocFixed(Memory.MainMem.GetStartAddr() + l.GetMaxAddr(), 0xFFFFED - l.GetMaxAddr());
 		thread.SetEntry(l.GetEntry() - Memory.MainMem.GetStartAddr());
 	break;

@ -314,12 +314,12 @@ void Emulator::Load()
 		m_ppu_callback_thr = &GetCPU().AddThread(CPU_THREAD_PPU);

 		thread.SetEntry(l.GetEntry());
-		Memory.StackMem.Alloc(0x1000);
+		Memory.StackMem.AllocAlign(0x1000);
 		thread.InitStack();
 		thread.AddArgv(m_elf_path);
 		//thread.AddArgv("-emu");

-		m_rsx_callback = Memory.MainMem.Alloc(4 * 4) + 4;
+		m_rsx_callback = Memory.MainMem.AllocAlign(4 * 4) + 4;
 		Memory.Write32(m_rsx_callback - 4, m_rsx_callback);

 		mem32_ptr_t callback_data(m_rsx_callback);
@ -327,7 +327,7 @@ void Emulator::Load()
 		callback_data += SC(2);
 		callback_data += BCLR(0x10 | 0x04, 0, 0, 0);

-		m_ppu_thr_exit = Memory.MainMem.Alloc(4 * 4);
+		m_ppu_thr_exit = Memory.MainMem.AllocAlign(4 * 4);

 		mem32_ptr_t ppu_thr_exit_data(m_ppu_thr_exit);
 		ppu_thr_exit_data += ADDI(3, 0, 0);
--- a/rpcs3/Emu/event.h
+++ b/rpcs3/Emu/event.h
@ -1,61 +1,62 @@
-#pragma once
-
-struct sys_event_flag_attr
-{
-	u32 protocol;
-	u32 pshared;
-	u64 ipc_key;
-	int flags;
-	int type;
-	char name[8];
-};
-
-struct event_flag
-{
-	sys_event_flag_attr attr;
-	u64 pattern;
-
-	event_flag(u64 pattern, sys_event_flag_attr attr)
-		: pattern(pattern)
-		, attr(attr)
-	{
-	}
-};
-
-struct sys_event_queue_attr
-{
-	u32 attr_protocol;
-	int type;
-	char name[8];
-};
-
-struct sys_event_data
-{
-	u64 source;
-	u64 data1;
-	u64 data2;
-	u64 data3;
-};
-
-struct EventQueue;
-
-struct EventPort
-{
-	u64 name;
-	u64 data1;
-	u64 data2;
-	u64 data3;
-	bool has_data;
-	CPUThread* thread;
-	EventQueue* queue[127];
-	int pos;
-};
-
-struct EventQueue
-{
-	EventPort* ports[127];
-	int size;
-	int pos;
-	int type;
-	char name[8];
+#pragma once
+
+enum EventQueueType
+{
+	SYS_PPU_QUEUE = 1,
+	SYS_SPU_QUEUE = 2,
+};
+
+struct sys_event_queue_attr
+{
+	be_t<u32> protocol; // SYS_SYNC_PRIORITY or SYS_SYNC_FIFO
+	be_t<int> type;
+	union
+	{
+		char name[8];
+		u64 name_u64;
+	};
+};
+
+struct sys_event_data
+{
+	be_t<u64> source;
+	be_t<u64> data1;
+	be_t<u64> data2;
+	be_t<u64> data3;
+};
+
+struct EventQueue;
+
+struct EventPort
+{
+	u64 name;
+	u64 data1;
+	u64 data2;
+	u64 data3;
+	bool has_data;
+	CPUThread* thread;
+	EventQueue* queue[127];
+	int pos;
+};
+
+struct EventQueue
+{
+	EventPort* ports[127];
+	int size;
+	int pos;
+
+	u32 m_protocol;	
+	int m_type;
+	u64 m_name;
+	u64 m_key;
+
+	EventQueue(u32 protocol, int type, u64 name, u64 key, int size)
+		: m_type(type)
+		, m_protocol(protocol)
+		, m_name(name)
+		, m_key(key)
+		, size(size)
+		, pos(0)
+	{
+	}
 };
--- a/rpcs3/Loader/ELF32.cpp
+++ b/rpcs3/Loader/ELF32.cpp
@ -181,9 +181,9 @@ bool ELF32Loader::LoadPhdrData(u64 _offset)

 			switch(machine)
 			{
-			case MACHINE_SPU: Memory.MainMem.Alloc(phdr_arr[i].p_vaddr + offset, phdr_arr[i].p_memsz); break;
-			case MACHINE_MIPS: Memory.PSPMemory.RAM.Alloc(phdr_arr[i].p_vaddr + offset, phdr_arr[i].p_memsz); break;
-			case MACHINE_ARM: Memory.PSVMemory.RAM.Alloc(phdr_arr[i].p_vaddr + offset, phdr_arr[i].p_memsz); break;
+			case MACHINE_SPU: Memory.MainMem.AllocFixed(phdr_arr[i].p_vaddr + offset, phdr_arr[i].p_memsz); break;
+			case MACHINE_MIPS: Memory.PSPMemory.RAM.AllocFixed(phdr_arr[i].p_vaddr + offset, phdr_arr[i].p_memsz); break;
+			case MACHINE_ARM: Memory.PSVMemory.RAM.AllocFixed(phdr_arr[i].p_vaddr + offset, phdr_arr[i].p_memsz); break;

 			default:
 				continue;
--- a/rpcs3/Loader/ELF64.cpp
+++ b/rpcs3/Loader/ELF64.cpp
@ -240,7 +240,7 @@ bool ELF64Loader::LoadPhdrData(u64 offset)
 			case 0x00000001: //LOAD
 				if(phdr_arr[i].p_memsz)
 				{
-					Memory.MainMem.Alloc(offset + phdr_arr[i].p_vaddr, phdr_arr[i].p_memsz);
+					Memory.MainMem.AllocFixed(offset + phdr_arr[i].p_vaddr, phdr_arr[i].p_memsz);

 					if(phdr_arr[i].p_filesz)
 					{
@ -358,8 +358,8 @@ bool ELF64Loader::LoadPhdrData(u64 offset)
 						ConLog.Write("*** text: 0x%x", stub.s_text);
 #endif
 						static const u32 section = 4 * 3;
-						u64 tbl = Memory.MainMem.Alloc(stub.s_imports * 4 * 2);
-						u64 dst = Memory.MainMem.Alloc(stub.s_imports * section);
+						u64 tbl = Memory.MainMem.AllocAlign(stub.s_imports * 4 * 2);
+						u64 dst = Memory.MainMem.AllocAlign(stub.s_imports * section);

 						for(u32 i=0; i<stub.s_imports; ++i)
 						{
--- a/rpcs3/Loader/Loader.h
+++ b/rpcs3/Loader/Loader.h
@ -197,7 +197,7 @@ public:

 	u32 GetEntry() { return entry; }
 	u32 GetMinAddr() { return min_addr; }
-	u32 GetMaxAddr() { return min_addr; }
+	u32 GetMaxAddr() { return max_addr; }
 };

 class Loader : public LoaderBase
--- a/rpcs3/rpcs3.vcxproj
+++ b/rpcs3/rpcs3.vcxproj
@ -69,20 +69,20 @@
  </ImportGroup>
  <PropertyGroup Label="UserMacros" />
  <PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
-    <IncludePath>.\;..\wxWidgets\include;..\SDL-1.3.0-5538\include;..\SDL_image-1.2.10;..\pthreads-2.8.0;..\;$(IncludePath)</IncludePath>
+    <IncludePath>.\;..\wxWidgets\include;..\SDL-1.3.0-5538\include;..\SDL_image-1.2.10;..\pthreads-2.8.0;..\;..\..\ffmpeg;$(IncludePath)</IncludePath>
    <OutDir>$(SolutionDir)bin\</OutDir>
    <LibraryPath>..\libs\$(Configuration)\;$(LibraryPath)</LibraryPath>
    <TargetName>$(ProjectName)-$(PlatformShortName)-dbg</TargetName>
  </PropertyGroup>
  <PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">
-    <IncludePath>.\;..\wxWidgets\include;..\SDL-1.3.0-5538\include;..\SDL_image-1.2.10;..\pthreads-2.8.0;..\;$(IncludePath)</IncludePath>
+    <IncludePath>.\;..\wxWidgets\include;..\SDL-1.3.0-5538\include;..\SDL_image-1.2.10;..\pthreads-2.8.0;..\;..\..\ffmpeg;$(IncludePath)</IncludePath>
    <OutDir>$(SolutionDir)bin\</OutDir>
    <LibraryPath>..\libs\$(Configuration)\;$(LibraryPath)</LibraryPath>
    <TargetName>$(ProjectName)-$(PlatformShortName)-dbg</TargetName>
  </PropertyGroup>
  <PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
    <LinkIncremental>false</LinkIncremental>
-    <IncludePath>.\;..\wxWidgets\include;..\SDL-1.3.0-5538\include;..\SDL_image-1.2.10;..\pthreads-2.8.0;..\;$(IncludePath)</IncludePath>
+    <IncludePath>.\;..\wxWidgets\include;..\SDL-1.3.0-5538\include;..\SDL_image-1.2.10;..\pthreads-2.8.0;..\;..\..\ffmpeg;$(IncludePath)</IncludePath>
    <OutDir>$(SolutionDir)bin\</OutDir>
    <LibraryPath>..\libs\$(Configuration)\;$(LibraryPath)</LibraryPath>
    <LinkIncremental>false</LinkIncremental>
@ -91,7 +91,7 @@
  </PropertyGroup>
  <PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'">
    <LinkIncremental>false</LinkIncremental>
-    <IncludePath>.\;..\wxWidgets\include;..\SDL-1.3.0-5538\include;..\SDL_image-1.2.10;..\pthreads-2.8.0;..\;$(IncludePath)</IncludePath>
+    <IncludePath>.\;..\wxWidgets\include;..\SDL-1.3.0-5538\include;..\SDL_image-1.2.10;..\pthreads-2.8.0;..\;..\..\ffmpeg;$(IncludePath)</IncludePath>
    <OutDir>$(SolutionDir)bin\</OutDir>
    <LibraryPath>..\libs\$(Configuration)\;$(LibraryPath)</LibraryPath>
    <LinkIncremental>false</LinkIncremental>
@ -205,6 +205,7 @@
      <PrecompiledHeader Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">NotUsing</PrecompiledHeader>
      <PrecompiledHeader Condition="'$(Configuration)|$(Platform)'=='Release|x64'">NotUsing</PrecompiledHeader>
    </ClCompile>
+    <ClCompile Include="..\Utilities\SMutex.cpp" />
    <ClCompile Include="..\Utilities\Thread.cpp" />
    <ClCompile Include="AppConnector.cpp" />
    <ClCompile Include="Emu\Audio\AudioManager.cpp" />
@ -247,6 +248,7 @@
    <ClCompile Include="Emu\SysCalls\FuncList.cpp" />
    <ClCompile Include="Emu\SysCalls\lv2\SC_Condition.cpp" />
    <ClCompile Include="Emu\SysCalls\lv2\SC_Event.cpp" />
+    <ClCompile Include="Emu\SysCalls\lv2\SC_Event_flag.cpp" />
    <ClCompile Include="Emu\SysCalls\lv2\SC_FileSystem.cpp" />
    <ClCompile Include="Emu\SysCalls\lv2\SC_GCM.cpp" />
    <ClCompile Include="Emu\SysCalls\lv2\SC_Heap.cpp" />
@ -324,6 +326,7 @@
    <ClInclude Include="..\Utilities\BEType.h" />
    <ClInclude Include="..\Utilities\IdManager.h" />
    <ClInclude Include="..\Utilities\MTProgressDialog.h" />
+    <ClInclude Include="..\Utilities\SMutex.h" />
    <ClInclude Include="..\Utilities\Thread.h" />
    <ClInclude Include="..\Utilities\Timer.h" />
    <ClInclude Include="Emu\Audio\AudioManager.h" />
--- a/rpcs3/rpcs3.vcxproj.filters
+++ b/rpcs3/rpcs3.vcxproj.filters
@ -379,6 +379,12 @@
    <ClCompile Include="Emu\FS\vfsLocalDir.cpp">
      <Filter>Emu\FS</Filter>
    </ClCompile>
+    <ClCompile Include="..\Utilities\SMutex.cpp">
+      <Filter>Utilities</Filter>
+    </ClCompile>
+    <ClCompile Include="Emu\SysCalls\lv2\SC_Event_flag.cpp">
+      <Filter>Emu\SysCalls\lv2</Filter>
+    </ClCompile>
  </ItemGroup>
  <ItemGroup>
    <ResourceCompile Include="rpcs3.rc" />
@ -558,5 +564,8 @@
    <ClInclude Include="Emu\Audio\AudioManager.h">
      <Filter>Include</Filter>
    </ClInclude>
+    <ClInclude Include="..\Utilities\SMutex.h">
+      <Filter>Utilities</Filter>
+    </ClInclude>
  </ItemGroup>
 </Project>