[InlineCost]Account for switch instructons when the switch condition could be simplified as a result of inlines.

[kazu]

See the comment about isa_and_present above.

llvm/lib/Analysis/InlineCost.cpp

859–860

May I suggest something like this?

if (isa_and_present<ConstantInt>(SimplifiedValues.lookup(SI->getCondition())))

This is for consistency with the BranchInst handling just a few lines above.

isa_and_present is a successor of isa_and_nonnull, which I *think* will be deprecated.

[mingmingl]

address feedback.

llvm/lib/Analysis/InlineCost.cpp

for (auto &BB : F) {
       for (auto &I : BB) {
         if (BranchInst *BI = dyn_cast<BranchInst>(&I)) {
           // Count a conditional branch as savings if it becomes unconditional.
           if (BI->isConditional() &&
               isa_and_nonnull<ConstantInt>(
                   SimplifiedValues.lookup(BI->getCondition()))) {
             CurrentSavings += InstrCost;
           }
+        } else if (SwitchInst *SI = dyn_cast<SwitchInst>(&I)) {
+          if (isa_and_present<ConstantInt>(SimplifiedValues.lookup(SI->getCondition())))
+            CurrentSavings += InstrCost;
         } else if (Value *V = dyn_cast<Value>(&I)) {
           // Count an instruction as savings if we can fold it.
           if (SimplifiedValues.count(V)) {
             CurrentSavings += InstrCost;
           }
         }
       }