[ARM32] Adds unsigned compares.

Superscedes PR #1131.
Optimizes cmpi a bit.

Source/Core/Core/PowerPC/JitArm32/Jit.h

@@ -178,7 +178,9 @@ public:
 	void subfic(UGeckoInstruction _inst);
 	void cntlzwx(UGeckoInstruction _inst);
 	void cmp (UGeckoInstruction _inst);
+	void cmpl(UGeckoInstruction _inst);
 	void cmpi(UGeckoInstruction _inst);
+	void cmpli(UGeckoInstruction _inst);
 	void negx(UGeckoInstruction _inst);
 	void mulhwux(UGeckoInstruction _inst);
 	void rlwimix(UGeckoInstruction _inst);

Source/Core/Core/PowerPC/JitArm32/JitArm_Integer.cpp

@@ -825,10 +825,34 @@ void JitArm::cmp (UGeckoInstruction inst)
 
 	gpr.Unlock(rA);
 }
+
+void JitArm::cmpl(UGeckoInstruction inst)
+{
+	INSTRUCTION_START
+	JITDISABLE(bJITIntegerOff);
+
+	int crf = inst.CRFD;
+	u32 a = inst.RA, b = inst.RB;
+
+	if (gpr.IsImm(a) && gpr.IsImm(b))
+	{
+		ComputeRC(gpr.GetImm(a) - gpr.GetImm(b), crf);
+		return;
+	}
+	else if (gpr.IsImm(b) && !gpr.GetImm(b))
+	{
+		ComputeRC(gpr.R(a), crf);
+		return;
+	}
+
+	FALLBACK_IF(true);
+}
+
 void JitArm::cmpi(UGeckoInstruction inst)
 {
 	INSTRUCTION_START
 	JITDISABLE(bJITIntegerOff);
+
 	u32 a = inst.RA;
 	int crf = inst.CRFD;
 	if (gpr.IsImm(a))
@@ -838,10 +862,15 @@ void JitArm::cmpi(UGeckoInstruction inst)
 	}
 	ARMReg rA = gpr.GetReg();
 	ARMReg RA = gpr.R(a);
+	bool negated = false;
+	Operand2 off;
 
-	if (inst.SIMM_16 >= 0 && inst.SIMM_16 < 256)
+	if (TryMakeOperand2_AllowNegation(inst.SIMM_16, off, &negated))
 	{
-		SUB(rA, RA, inst.SIMM_16);
+		if (negated)
+			ADD(rA, RA, off);
+		else
+			SUB(rA, RA, off);
 	}
 	else
 	{
@@ -853,6 +882,28 @@ void JitArm::cmpi(UGeckoInstruction inst)
 	gpr.Unlock(rA);
 }
 
+void JitArm::cmpli(UGeckoInstruction inst)
+{
+	INSTRUCTION_START
+	JITDISABLE(bJITIntegerOff);
+	u32 a = inst.RA;
+	int crf = inst.CRFD;
+
+	if (gpr.IsImm(a))
+	{
+		ComputeRC(gpr.GetImm(a) - inst.UIMM, crf);
+		return;
+	}
+
+	if (!inst.UIMM)
+	{
+		ComputeRC(gpr.R(a), crf);
+		return;
+	}
+
+	FALLBACK_IF(true);
+}
+
 void JitArm::negx(UGeckoInstruction inst)
 {
 	INSTRUCTION_START

Source/Core/Core/PowerPC/JitArm32/JitArm_Tables.cpp

@@ -47,7 +47,7 @@ static GekkoOPTemplate primarytable[] =
 
 	{7,  &JitArm::arith},                 //"mulli",    OPTYPE_INTEGER, FL_OUT_D | FL_IN_A | FL_RC_BIT, 2}},
 	{8,  &JitArm::subfic},                //"subfic",   OPTYPE_INTEGER, FL_OUT_D | FL_IN_A | FL_SET_CA}},
-	{10, &JitArm::FallBackToInterpreter}, //"cmpli",    OPTYPE_INTEGER, FL_IN_A | FL_SET_CRn}},
+	{10, &JitArm::cmpli},                 //"cmpli",    OPTYPE_INTEGER, FL_IN_A | FL_SET_CRn}},
 	{11, &JitArm::cmpi},                  //"cmpi",     OPTYPE_INTEGER, FL_IN_A | FL_SET_CRn}},
 	{12, &JitArm::arith},                 //"addic",    OPTYPE_INTEGER, FL_OUT_D | FL_IN_A | FL_SET_CA}},
 	{13, &JitArm::arith},                 //"addic_rc", OPTYPE_INTEGER, FL_OUT_D | FL_IN_A | FL_SET_CR0}},
@@ -190,7 +190,7 @@ static GekkoOPTemplate table31[] =
 	{476, &JitArm::arith},                  //"nandx",  OPTYPE_INTEGER, FL_OUT_A | FL_IN_SB | FL_RC_BIT}},
 	{284, &JitArm::arith},                  //"eqvx",   OPTYPE_INTEGER, FL_OUT_A | FL_IN_SB | FL_RC_BIT}},
 	{0,   &JitArm::cmp},                    //"cmp",    OPTYPE_INTEGER, FL_IN_AB | FL_SET_CRn}},
-	{32,  &JitArm::FallBackToInterpreter},  //"cmpl",   OPTYPE_INTEGER, FL_IN_AB | FL_SET_CRn}},
+	{32,  &JitArm::cmpl},                   //"cmpl",   OPTYPE_INTEGER, FL_IN_AB | FL_SET_CRn}},
 	{26,  &JitArm::cntlzwx},                //"cntlzwx",OPTYPE_INTEGER, FL_OUT_A | FL_IN_S | FL_RC_BIT}},
 	{922, &JitArm::extshx},                 //"extshx", OPTYPE_INTEGER, FL_OUT_A | FL_IN_S | FL_RC_BIT}},
 	{954, &JitArm::extsbx},                 //"extsbx", OPTYPE_INTEGER, FL_OUT_A | FL_IN_S | FL_RC_BIT}},