[OpenMP][OpenMPIRBuilder] Fix static code analysis concerns with uninitialized variables

[Manna]

Summary

computeHeuristicUnrollFactor(): Fix static analyzer warnings with uninitialized variables warnings - NFCI

[aaron.ballman]

The static analyzer is broken and giving a false positive report. llvm::ApproximateLoopSize() always assigns to those byref arguments. I'm not certain there's much value from forcing an initialization here, but if LLVM folks want the changes, I wouldn't block them.

[jdoerfert]

LG, still probably good to improve the analyzer.

llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp

for (Instruction &I : *BB) {
 
       if (auto *Alloca = dyn_cast<AllocaInst>(Ptr)) {
         if (Alloca->getParent() == &F->getEntryBlock())
           EphValues.insert(&I);
       }
     }
   }
 
-  unsigned NumInlineCandidates;
-  bool NotDuplicatable;
-  bool Convergent;
+  unsigned NumInlineCandidates = 0;
+  bool NotDuplicatable = false;
+  bool Convergent = false;
   InstructionCost LoopSizeIC =
       ApproximateLoopSize(L, NumInlineCandidates, NotDuplicatable, Convergent,
                           TTI, EphValues, UP.BEInsns);
   LLVM_DEBUG(dbgs() << "Estimated loop size is " << LoopSizeIC << "\n");
 
   // Loop is not unrollable if the loop contains certain instructions.
   if (NotDuplicatable || Convergent || !LoopSizeIC.isValid()) {
     LLVM_DEBUG(dbgs() << "Loop not considered unrollable\n");