De-duplicate the two implementations of ARMBaseInstrInfo::isProfitableToIfCvt() [NFC]

[rengolin]

Looks simple enough. I'm guessing we have enough tests to make sure nothing breaks with this change (though, I don't expect anything to break).

lib/Target/ARM/ARMBaseInstrInfo.cpp

if (!Pred->empty()) {
             if (P == ARMCC::AL && CmpMI->getOperand(1).getImm() == 0 &&
                 isARMLowRegister(Reg))
               return false;
           }
         }
       }
     }
   }
-
-  // Attempt to estimate the relative costs of predication versus branching.
-  // Here we scale up each component of UnpredCost to avoid precision issue when
-  // scaling NumCycles by Probability.
-  const unsigned ScalingUpFactor = 1024;
-  unsigned UnpredCost = Probability.scale(NumCycles * ScalingUpFactor);
-  UnpredCost += ScalingUpFactor; // The branch itself
-  UnpredCost += Subtarget.getMispredictionPenalty() * ScalingUpFactor / 10;
-
-  return (NumCycles + ExtraPredCycles) * ScalingUpFactor <= UnpredCost;
+  return isProfitableToIfCvt(MBB, NumCycles, ExtraPredCycles,
+                             MBB, 0, 0, Probability);
 }
 
 bool ARMBaseInstrInfo::
-isProfitableToIfCvt(MachineBasicBlock &TMBB,
+isProfitableToIfCvt(MachineBasicBlock &,
                     unsigned TCycles, unsigned TExtra,
-                    MachineBasicBlock &FMBB,
+                    MachineBasicBlock &,
                     unsigned FCycles, unsigned FExtra,
                     BranchProbability Probability) const {
-  if (!TCycles || !FCycles)
+  if (!TCycles)
     return false;
 
   // Attempt to estimate the relative costs of predication versus branching.
   // Here we scale up each component of UnpredCost to avoid precision issue when
   // scaling TCycles/FCycles by Probability.
   const unsigned ScalingUpFactor = 1024;
   unsigned TUnpredCost = Probability.scale(TCycles * ScalingUpFactor);
   unsigned FUnpredCost =