diff --git a/Source/Core/Common/Src/ABI.cpp b/Source/Core/Common/Src/ABI.cpp
index e45b93fbe2..2c4d2ceab9 100644
--- a/Source/Core/Common/Src/ABI.cpp
+++ b/Source/Core/Common/Src/ABI.cpp
@@ -29,6 +29,14 @@ void ABI_CallFunctionR(void *func, X64Reg reg1) {
 	ADD(32, R(ESP), Imm8(4));
 }
 
+void ABI_CallFunctionRR(void *func, Gen::X64Reg reg1, Gen::X64Reg reg2)
+{
+	PUSH(32, R(reg2));
+	PUSH(32, R(reg1));
+	CALL(func);
+	ADD(32, R(ESP), Imm8(8));
+}
+
 void ABI_CallFunctionAC(void *func, const Gen::OpArg &arg1, u32 param2)
 {
 	PUSH(32, arg1);
@@ -74,6 +82,15 @@ void ABI_CallFunctionR(void *func, X64Reg reg1) {
 	CALL(func);
 }
 
+// Pass a register as a paremeter.
+void ABI_CallFunctionRR(void *func, X64Reg reg1, X64Reg reg2) {
+	if (reg1 != ABI_PARAM1)
+		MOV(32, R(ABI_PARAM1), R(reg1));
+	if (reg2 != ABI_PARAM2)
+		MOV(32, R(ABI_PARAM2), R(reg2));
+	CALL(func);
+}
+
 void ABI_CallFunctionAC(void *func, const Gen::OpArg &arg1, u32 param2)
 {
 	if (!arg1.IsSimpleReg(ABI_PARAM1))
diff --git a/Source/Core/Common/Src/ABI.h b/Source/Core/Common/Src/ABI.h
index 3386913834..632df334ef 100644
--- a/Source/Core/Common/Src/ABI.h
+++ b/Source/Core/Common/Src/ABI.h
@@ -48,6 +48,11 @@
 #ifdef _M_IX86
 // 32 bit calling convention, shared by all
 
+// 32-bit don't pass parameters in regs, but these are convenient to have anyway when we have to
+// choose regs to put stuff in.
+#define ABI_PARAM1 RCX
+#define ABI_PARAM2 RDX
+
 // There are no ABI_PARAM* here, since args are pushed.
 
 // === 32-bit bog standard cdecl, shared between linux and windows ============================
@@ -85,6 +90,7 @@ void ABI_CallFunctionAC(void *func, const Gen::OpArg &arg1, u32 param2);
 
 // Pass a register as a paremeter.
 void ABI_CallFunctionR(void *func, Gen::X64Reg reg1);
+void ABI_CallFunctionRR(void *func, Gen::X64Reg reg1, Gen::X64Reg reg2);
 
 void ABI_PushAllCalleeSavedRegsAndAdjustStack();
 void ABI_PopAllCalleeSavedRegsAndAdjustStack();
diff --git a/Source/Core/Core/Src/PowerPC/Jit64/Jit_LoadStore.cpp b/Source/Core/Core/Src/PowerPC/Jit64/Jit_LoadStore.cpp
index 8c3fc9a425..1c93d25c4c 100644
--- a/Source/Core/Core/Src/PowerPC/Jit64/Jit_LoadStore.cpp
+++ b/Source/Core/Core/Src/PowerPC/Jit64/Jit_LoadStore.cpp
@@ -87,10 +87,10 @@ namespace Jit64
 			gpr.LoadToX64(d, true, true);
 		else 
 			gpr.LoadToX64(d, false, true);
-		MOV(32, R(ECX), gpr.R(b));
+		MOV(32, R(ABI_PARAM1), gpr.R(b));
 		if (a)
-			ADD(32, R(ECX), gpr.R(a));
-		SafeLoadRegToEAX(ECX, 8, 0);
+			ADD(32, R(ABI_PARAM1), gpr.R(a));
+		SafeLoadRegToEAX(ABI_PARAM1, 8, 0);
 		MOV(32, gpr.R(d), R(EAX));
 		gpr.UnlockAll();
 	}
@@ -143,8 +143,8 @@ namespace Jit64
 			// Safe and boring
 			gpr.Flush(FLUSH_VOLATILE);
 			gpr.Lock(d, a);
-			MOV(32, R(ECX), gpr.R(a));
-			SafeLoadRegToEAX(ECX, accessSize, offset);
+			MOV(32, R(ABI_PARAM1), gpr.R(a));
+			SafeLoadRegToEAX(ABI_PARAM1, accessSize, offset);
 			gpr.LoadToX64(d, false, true);
 			MOV(32, gpr.R(d), R(EAX));
 			gpr.UnlockAll();
@@ -184,7 +184,7 @@ namespace Jit64
 		gpr.Flush(FLUSH_VOLATILE);
 		gpr.Lock(d, a);
 		
-		MOV(32,R(ECX), gpr.R(a));
+		MOV(32, R(ABI_PARAM1), gpr.R(a));
 #ifdef _M_X64
 		if (!jo.noAssumeFPLoadFromMem)
 		{
@@ -197,7 +197,7 @@ namespace Jit64
 		else
 #endif
 		{
-			SafeLoadRegToEAX(ECX, 32, offset);
+			SafeLoadRegToEAX(ABI_PARAM1, 32, offset);
 		}
 
 		MOV(32, M(&temp32), R(EAX));
@@ -271,21 +271,21 @@ namespace Jit64
 			gpr.Flush(FLUSH_VOLATILE);
 			gpr.Lock(a);
 			fpr.Lock(s);
-			MOV(32,R(EDX),gpr.R(a));
+			MOV(32, R(ABI_PARAM2), gpr.R(a));
 			if (offset)
-				ADD(32,R(EDX),Imm32((u32)offset));
+				ADD(32, R(ABI_PARAM2), Imm32((u32)offset));
 			if (update && offset)
 			{
-				MOV(32,gpr.R(a),R(EDX));
+				MOV(32, gpr.R(a), R(ABI_PARAM2));
 			}
-			CVTSD2SS(XMM0,fpr.R(s));
-			MOVSS(M(&temp32),XMM0);
-			MOV(32,R(ECX),M(&temp32));
+			CVTSD2SS(XMM0, fpr.R(s));
+			MOVSS(M(&temp32), XMM0);
+			MOV(32, R(ABI_PARAM1), M(&temp32));
 
-			TEST(32,R(EDX),Imm32(0x0C000000));
+			TEST(32, R(ABI_PARAM2), Imm32(0x0C000000));
 			FixupBranch argh = J_CC(CC_NZ);
-			BSWAP(32,ECX);
-			MOV(32, MComplex(RBX, EDX, SCALE_1, 0), R(ECX));
+			BSWAP(32, ABI_PARAM1);
+			MOV(32, MComplex(RBX, ABI_PARAM2, SCALE_1, 0), R(ABI_PARAM1));
 			FixupBranch arg2 = J();
 			SetJumpTarget(argh);
 			CALL((void *)&Memory::Write_U32); 
@@ -358,7 +358,7 @@ namespace Jit64
 			case 38: accessSize = 8; break;  //stb
 			default: _assert_msg_(DYNA_REC, 0, "AWETKLJASDLKF"); return;
 			}
-
+/*
 			if (gpr.R(a).IsImm() && !update)
 			{
 				u32 addr = (u32)gpr.R(a).offset;
@@ -424,48 +424,41 @@ namespace Jit64
 					ADD(32, gpr.R(a), Imm32(offset));
 				return;
 			}
-
+*/
 			//Still here? Do regular path.
 			gpr.Lock(s, a);
-			MOV(32, R(EDX), gpr.R(a));
-			MOV(32, R(ECX), gpr.R(s));
+			MOV(32, R(ABI_PARAM2), gpr.R(a));
+			MOV(32, R(ABI_PARAM1), gpr.R(s));
 			if (offset)
-				ADD(32, R(EDX), Imm32((u32)offset));
+				ADD(32, R(ABI_PARAM2), Imm32((u32)offset));
 			if (update && offset)
 			{
-				MOV(32, gpr.R(a), R(EDX));
+				MOV(32, gpr.R(a), R(ABI_PARAM2));
 			}
-			TEST(32, R(EDX), Imm32(0x0C000000));
+			TEST(32, R(ABI_PARAM2), Imm32(0x0C000000));
 			FixupBranch argh = J_CC(CC_NZ);
 			if (accessSize == 32)
-				BSWAP(32, ECX);
+				BSWAP(32, ABI_PARAM1);
 			else if (accessSize == 16)
 			{
 				// TODO(ector): xchg ch, cl?
-				BSWAP(32, ECX);
-				SHR(32, R(ECX), Imm8(16));
+				BSWAP(32, ABI_PARAM1);
+				SHR(32, R(ABI_PARAM1), Imm8(16));
 			}
 #ifdef _M_X64
-			MOV(accessSize, MComplex(RBX, EDX, SCALE_1, 0), R(ECX));
+			MOV(accessSize, MComplex(RBX, ABI_PARAM2, SCALE_1, 0), R(ABI_PARAM1));
 #else
-			AND(32, R(EDX), Imm32(Memory::MEMVIEW32_MASK));
-			MOV(accessSize, MDisp(EDX, (u32)Memory::base), R(ECX));
+			AND(32, R(ABI_PARAM2), Imm32(Memory::MEMVIEW32_MASK));
+			MOV(accessSize, MDisp(ABI_PARAM2, (u32)Memory::base), R(ABI_PARAM1));
 #endif
 			FixupBranch arg2 = J();
 			SetJumpTarget(argh);
-#ifdef _M_IX86
-			PUSH(EDX);
-			PUSH(ECX);
-#endif
 			switch (accessSize)
 			{
-			case 32: CALL((void *)&Memory::Write_U32); break;
-			case 16: CALL((void *)&Memory::Write_U16); break;
-			case 8:  CALL((void *)&Memory::Write_U8); break;
+			case 32: ABI_CallFunctionRR((void *)&Memory::Write_U32, ABI_PARAM1, ABI_PARAM2); break;
+			case 16: ABI_CallFunctionRR((void *)&Memory::Write_U16, ABI_PARAM1, ABI_PARAM2); break;
+			case 8:  ABI_CallFunctionRR((void *)&Memory::Write_U8, ABI_PARAM1, ABI_PARAM2); break;
 			}
-#ifdef _M_IX86
-			ADD(32, R(ESP), Imm8(8));
-#endif
 			SetJumpTarget(arg2);
 			gpr.UnlockAll();
 		}
diff --git a/Source/Core/Core/Src/PowerPC/Jit64/Jit_LoadStorePaired.cpp b/Source/Core/Core/Src/PowerPC/Jit64/Jit_LoadStorePaired.cpp
index 94128419da..7bfd331e78 100644
--- a/Source/Core/Core/Src/PowerPC/Jit64/Jit_LoadStorePaired.cpp
+++ b/Source/Core/Core/Src/PowerPC/Jit64/Jit_LoadStorePaired.cpp
@@ -26,6 +26,7 @@
 #include "../../HW/Memmap.h"
 #include "../PPCTables.h"
 #include "x64Emitter.h"
+#include "ABI.h"
 
 #include "Jit.h"
 #include "JitCache.h"
@@ -111,7 +112,7 @@ void psq_st(UGeckoInstruction inst)
 	const EQuantizeType stType = static_cast<EQuantizeType>(gqr.ST_TYPE);
 	int stScale = gqr.ST_SCALE;
 	bool update = inst.OPCD == 61;
-	if (!inst.RA || inst.W)
+	if (update || !inst.RA || inst.W)
 	{
 	//	PanicAlert(inst.RA ? "W" : "inst");
 		Default(inst);
@@ -129,25 +130,25 @@ void psq_st(UGeckoInstruction inst)
 		fpr.Lock(s);
 		if (update)
 			gpr.LoadToX64(a, true, true);
-		MOV(32, R(EDX), gpr.R(a));
+		MOV(32, R(ABI_PARAM2), gpr.R(a));
 		if (offset)
-			ADD(32, R(EDX), Imm32((u32)offset));
-		TEST(32, R(EDX), Imm32(0x0C000000));
+			ADD(32, R(ABI_PARAM2), Imm32((u32)offset));
+		TEST(32, R(ABI_PARAM2), Imm32(0x0C000000));
 		if (update && offset)
-			MOV(32, gpr.R(a), R(EDX));
+			MOV(32, gpr.R(a), R(ABI_PARAM2));
 		CVTPD2PS(XMM0, fpr.R(s));
 		SHUFPS(XMM0, R(XMM0), 1);
 		MOVAPS(M(&temp64), XMM0);
-		MOV(64, R(ECX), M(&temp64));
+		MOV(64, R(ABI_PARAM1), M(&temp64));
 		FixupBranch argh = J_CC(CC_NZ);
-		BSWAP(64, ECX);
-		MOV(64, MComplex(RBX, EDX, SCALE_1, 0), R(ECX));
+		BSWAP(64, ABI_PARAM1);
+		MOV(64, MComplex(RBX, ABI_PARAM2, SCALE_1, 0), R(ABI_PARAM1));
 		FixupBranch arg2 = J();
 		SetJumpTarget(argh);
 		CALL((void *)&WriteDual32); 
 		SetJumpTarget(arg2);
 		if (update)
-			MOV(32, gpr.R(a), R(EDX));
+			MOV(32, gpr.R(a), R(ABI_PARAM2));
 		gpr.UnlockAll();
 		fpr.UnlockAll();
 	}
@@ -158,9 +159,9 @@ void psq_st(UGeckoInstruction inst)
 		fpr.Lock(s);
 		if (update)
 			gpr.LoadToX64(a, true, update);
-		MOV(32, R(EDX), gpr.R(a));
+		MOV(32, R(ABI_PARAM2), gpr.R(a));
 		if (offset)
-			ADD(32,R(EDX),Imm32((u32)offset));
+			ADD(32, R(ABI_PARAM2), Imm32((u32)offset));
 		MOVAPS(XMM0, fpr.R(s));
 		MOVDDUP(XMM1, M((void*)&m_quantizeTableD[stScale]));
 		MULPD(XMM0, R(XMM1));
@@ -168,16 +169,16 @@ void psq_st(UGeckoInstruction inst)
 		PACKSSDW(XMM0, R(XMM0));
 		PACKUSWB(XMM0, R(XMM0));
 		MOVAPS(M(&temp64), XMM0);
-		MOV(16, R(ECX), M(&temp64));
+		MOV(16, R(ABI_PARAM1), M(&temp64));
 #ifdef _M_X64
-		MOV(16, MComplex(RBX, RDX, SCALE_1, 0), R(ECX));
+		MOV(16, MComplex(RBX, RDX, SCALE_1, 0), R(ABI_PARAM1));
 #else
-		BSWAP(32, ECX);
-		SHR(32, R(ECX), Imm8(16));
+		BSWAP(32, ABI_PARAM1);
+		SHR(32, R(ABI_PARAM1), Imm8(16));
 		CALL(&Memory::Write_U16);
 #endif
 		if (update)
-			MOV(32, gpr.R(a), R(EDX));
+			MOV(32, gpr.R(a), R(ABI_PARAM2));
 		gpr.UnlockAll();
 		fpr.UnlockAll();
 	} 
@@ -187,9 +188,9 @@ void psq_st(UGeckoInstruction inst)
 		fpr.Lock(s);
 		if (update)
 			gpr.LoadToX64(a, true, update);
-		MOV(32, R(EDX), gpr.R(a));
+		MOV(32, R(ABI_PARAM2), gpr.R(a));
 		if (offset)
-			ADD(32,R(EDX),Imm32((u32)offset));
+			ADD(32, R(ABI_PARAM2), Imm32((u32)offset));
 		MOVAPS(XMM0, fpr.R(s));
 		MOVDDUP(XMM1, M((void*)&m_quantizeTableD[stScale]));
 		MULPD(XMM0, R(XMM1));
@@ -197,16 +198,16 @@ void psq_st(UGeckoInstruction inst)
 		CVTPD2DQ(XMM0, R(XMM0));
 		PACKSSDW(XMM0, R(XMM0));
 		MOVD_xmm(M(&temp64), XMM0);
-		MOV(32, R(ECX), M(&temp64));
+		MOV(32, R(ABI_PARAM1), M(&temp64));
 #ifdef _M_X64
-		BSWAP(32, ECX);
-		MOV(32, MComplex(RBX, RDX, SCALE_1, 0), R(ECX));
+		BSWAP(32, ABI_PARAM1);
+		MOV(32, MComplex(RBX, RDX, SCALE_1, 0), R(ABI_PARAM1));
 #else
-		BSWAP(32, ECX);
+		BSWAP(32, ABI_PARAM1);
 		CALL(&Memory::Write_U32);
 #endif
 		if (update)
-			MOV(32, gpr.R(a), R(EDX));
+			MOV(32, gpr.R(a), R(ABI_PARAM2));
 		gpr.UnlockAll();
 		fpr.UnlockAll();
 	}
@@ -231,14 +232,14 @@ void psq_l(UGeckoInstruction inst)
 	const UGQR gqr(rSPR(SPR_GQR0 + inst.I));
 	const EQuantizeType ldType = static_cast<EQuantizeType>(gqr.LD_TYPE);
 	int ldScale = gqr.LD_SCALE;
-	if (!inst.RA || inst.W)
+	bool update = inst.OPCD == 57;
+	if (update || !inst.RA || inst.W)
 	{
 		// 0 1 during load
 		//PanicAlert("ld:%i %i", ldType, (int)inst.W);
 		Default(inst);
 		return;
 	}
-	bool update = inst.OPCD == 57;
 	int offset = inst.SIMM_12;
 	//INT3();
 	switch (ldType) {
@@ -248,7 +249,7 @@ void psq_l(UGeckoInstruction inst)
 			gpr.LoadToX64(inst.RA);
 			MOV(64, R(RAX), MComplex(RBX, gpr.R(inst.RA).GetSimpleReg(), 1, offset));
 			BSWAP(64, RAX);
-			MOV(64, M(&psTemp[0]),R(RAX));
+			MOV(64, M(&psTemp[0]), R(RAX));
 			fpr.LoadToX64(inst.RS, false);
 			X64Reg r = fpr.R(inst.RS).GetSimpleReg();
 			CVTPS2PD(r, M(&psTemp[0]));