diff --git a/rpcs3/Emu/Cell/PPUThread.cpp b/rpcs3/Emu/Cell/PPUThread.cpp
index d8fb12c3ec..1d9d336a2a 100644
--- a/rpcs3/Emu/Cell/PPUThread.cpp
+++ b/rpcs3/Emu/Cell/PPUThread.cpp
@@ -671,7 +671,6 @@ void PPUThread::FastCall2(u32 addr, u32 rtoc)
 
 	m_status = Running;
 	PC = addr;
-	GPR[1] -= 0x70; // create stack frame reserved area
 	GPR[2] = rtoc;
 	LR = Emu.GetCPUThreadStop();
 	SetCurrentNamedThread(this);
@@ -681,7 +680,13 @@ void PPUThread::FastCall2(u32 addr, u32 rtoc)
 
 	m_status = old_status;
 	PC = old_PC;
-	GPR[1] = old_stack; // TODO: error check instead of blind assignment? GPR[1] shouldn't change
+
+	if (GPR[1] != old_stack) // GPR[1] shouldn't change
+	{
+		LOG_ERROR(PPU, "PPUThread::FastCall2(0x%x,0x%x): stack inconsistency (SP=0x%llx, old=0x%llx)", addr, rtoc, GPR[1], old_stack);
+		GPR[1] = old_stack;
+	}
+
 	GPR[2] = old_rtoc;
 	LR = old_LR;
 	SetCurrentNamedThread(old_thread);
diff --git a/rpcs3/Emu/Memory/vm.cpp b/rpcs3/Emu/Memory/vm.cpp
index dc32c017ac..c145c3bf57 100644
--- a/rpcs3/Emu/Memory/vm.cpp
+++ b/rpcs3/Emu/Memory/vm.cpp
@@ -628,27 +628,25 @@ namespace vm
 
 	u32 stack_push(CPUThread& CPU, u32 size, u32 align_v, u32& old_pos)
 	{
-		assert(align_v);
-
 		switch (CPU.GetType())
 		{
 		case CPU_THREAD_PPU:
 		{
-			PPUThread& PPU = static_cast<PPUThread&>(CPU);
+			PPUThread& context = static_cast<PPUThread&>(CPU);
 
-			old_pos = vm::cast(PPU.GPR[1], "SP");
-			PPU.GPR[1] -= align(size, 8); // room minimal possible size
-			PPU.GPR[1] &= ~(align_v - 1); // fix stack alignment
+			old_pos = vm::cast(context.GPR[1], "SP");
+			context.GPR[1] -= align(size, 8); // room minimal possible size
+			context.GPR[1] &= ~(align_v - 1); // fix stack alignment
 
-			if (PPU.GPR[1] < CPU.GetStackAddr())
+			if (context.GPR[1] < CPU.GetStackAddr())
 			{
-				// stack overflow
-				PPU.GPR[1] = old_pos;
+				LOG_ERROR(PPU, "vm::stack_push(0x%x,%d): stack overflow (SP=0x%llx, stack=*0x%x)", size, align_v, context.GPR[1], CPU.GetStackAddr());
+				context.GPR[1] = old_pos;
 				return 0;
 			}
 			else
 			{
-				return (u32)PPU.GPR[1];
+				return static_cast<u32>(context.GPR[1]);
 			}
 		}
 
@@ -669,7 +667,7 @@ namespace vm
 
 			if (context.SP < CPU.GetStackAddr())
 			{
-				// stack overflow
+				LOG_ERROR(ARMv7, "vm::stack_push(0x%x,%d): stack overflow (SP=0x%x, stack=*0x%x)", size, align_v, context.SP, CPU.GetStackAddr());
 				context.SP = old_pos;
 				return 0;
 			}
@@ -693,10 +691,14 @@ namespace vm
 		{
 		case CPU_THREAD_PPU:
 		{
-			PPUThread& PPU = static_cast<PPUThread&>(CPU);
+			PPUThread& context = static_cast<PPUThread&>(CPU);
 
-			assert(PPU.GPR[1] == addr);
-			PPU.GPR[1] = old_pos;
+			if (context.GPR[1] != addr)
+			{
+				LOG_ERROR(PPU, "vm::stack_pop(*0x%x,*0x%x): stack inconsistency (SP=0x%llx)", addr, old_pos, context.GPR[1]);
+			}
+
+			context.GPR[1] = old_pos;
 			return;
 		}
 
@@ -711,7 +713,11 @@ namespace vm
 		{
 			ARMv7Context& context = static_cast<ARMv7Thread&>(CPU).context;
 
-			assert(context.SP == addr);
+			if (context.SP != addr)
+			{
+				LOG_ERROR(ARMv7, "vm::stack_pop(*0x%x,*0x%x): stack inconsistency (SP=0x%x)", addr, old_pos, context.SP);
+			}
+
 			context.SP = old_pos;
 			return;
 		}
diff --git a/rpcs3/Emu/SysCalls/CB_FUNC.h b/rpcs3/Emu/SysCalls/CB_FUNC.h
index 2fb0a99a70..13203e43dc 100644
--- a/rpcs3/Emu/SysCalls/CB_FUNC.h
+++ b/rpcs3/Emu/SysCalls/CB_FUNC.h
@@ -132,10 +132,7 @@ namespace cb_detail
 	{
 		__forceinline static RT call(PPUThread& CPU, u32 pc, u32 rtoc, T... args)
 		{
-			const bool stack = _bind_func_args<0, 0, 0, T...>(CPU, args...);
-			if (stack) CPU.GPR[1] -= FIXED_STACK_FRAME_SIZE;
-			CPU.FastCall2(pc, rtoc);
-			if (stack) CPU.GPR[1] += FIXED_STACK_FRAME_SIZE;
+			_func_caller<void, T...>::call(CPU, pc, rtoc, args...);
 
 			static_assert(!std::is_pointer<RT>::value, "Invalid callback result type (pointer)");
 			static_assert(!std::is_reference<RT>::value, "Invalid callback result type (reference)");
@@ -154,7 +151,9 @@ namespace cb_detail
 		{
 			const bool stack = _bind_func_args<0, 0, 0, T...>(CPU, args...);
 			if (stack) CPU.GPR[1] -= FIXED_STACK_FRAME_SIZE;
+			CPU.GPR[1] -= 0x70; // create reserved area
 			CPU.FastCall2(pc, rtoc);
+			CPU.GPR[1] += 0x70;
 			if (stack) CPU.GPR[1] += FIXED_STACK_FRAME_SIZE;
 		}
 	};
diff --git a/rpcs3/Emu/SysCalls/Modules.cpp b/rpcs3/Emu/SysCalls/Modules.cpp
index c2cb8f2e0d..0e9f3561f9 100644
--- a/rpcs3/Emu/SysCalls/Modules.cpp
+++ b/rpcs3/Emu/SysCalls/Modules.cpp
@@ -106,12 +106,16 @@ void execute_ppu_func_by_index(PPUThread& CPU, u32 index)
 		{
 			// call LLE function if available
 
+			const auto data = vm::get_ptr<be_t<u32>>(func->lle_func.addr());
+			const u32 pc = data[0];
+			const u32 rtoc = data[1];
+
 			if (Ini.HLELogging.GetValue())
 			{
 				LOG_NOTICE(HLE, "LLE function called: %s", SysCalls::GetFuncName(func->id));
 			}
-
-			func->lle_func(CPU);
+			
+			CPU.FastCall2(pc, rtoc);
 
 			if (Ini.HLELogging.GetValue())
 			{