[clang][dataflow] Don't skip the entry block

[li.zhe.hua]

Summary

In D131646, we implement widen by taking the block state upon entry
into the loop and copying it to the block that represents the
back-edge in order to initialize the state for the "0th" iteration of
the loop. If the loop is the first thing after the entry block,
skipping the entry block means we don't have an opportunity to do this
initialization, so we want to analyze the entry block, even as it has
zero statements, to keep the logic consistent.

[li.zhe.hua]

Fix commit message

clang/lib/Analysis/FlowSensitive/TypeErasedDataflowAnalysis.cpp

std::function<void(const CFGStmt &,
                        const TypeErasedDataflowAnalysisState &)>
         PostVisitStmt) {
   PostOrderCFGView POV(&CFCtx.getCFG());
   ForwardDataflowWorklist Worklist(CFCtx.getCFG(), &POV);
 
   std::vector<llvm::Optional<TypeErasedDataflowAnalysisState>> BlockStates(
       CFCtx.getCFG().size(), llvm::None);
 
-  // The entry basic block doesn't contain statements so it can be skipped.
   const CFGBlock &Entry = CFCtx.getCFG().getEntry();
-  BlockStates[Entry.getBlockID()] = {Analysis.typeErasedInitialElement(),
-                                     InitEnv};
-  Worklist.enqueueSuccessors(&Entry);
+  Worklist.enqueueBlock(&Entry);
 
   // Bugs in lattices and transfer functions can prevent the analysis from
   // converging. To limit the damage (infinite loops) that these bugs can cause,
   // limit the number of iterations.
   // FIXME: Consider making the maximum number of iterations configurable.
   // FIXME: Consider restricting the number of backedges followed, rather than
   // iterations.
   // FIXME: Set up statistics (see llvm/ADT/Statistic.h) to count average number