[PM] Preserve the correct set of passes for GVN

[davide]

Summary

Not sure if this is just an oversight, or there's a real reason why we don't want to cache DominatorTree and GlobalsAA when the IR is mutated as the old PM does. Thanks!

[• dberlin]

LGTM

[davide]

Thanks for your review Daniel, r271934.

[chandlerc]

Just FYI for context:

lib/Transforms/Scalar/GVN.cpp

   // be less effective! We should fix memdep and basic-aa to not exhibit this
   // behavior, but until then don't change the order here.
   auto &AC = AM.getResult<AssumptionAnalysis>(F);
   auto &DT = AM.getResult<DominatorTreeAnalysis>(F);
   auto &TLI = AM.getResult<TargetLibraryAnalysis>(F);
   auto &AA = AM.getResult<AAManager>(F);
   auto &MemDep = AM.getResult<MemoryDependenceAnalysis>(F);
   bool Changed = runImpl(F, AC, DT, TLI, AA, &MemDep);
-  return Changed ? PreservedAnalyses::none() : PreservedAnalyses::all();
+  if (!Changed)
+    return PreservedAnalyses::all();
+  PreservedAnalyses PA;
+  PA.preserve<DominatorTreeAnalysis>();
+  PA.preserve<GlobalsAA>();
+  return PA;
 }
 
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
 void GVN::dump(DenseMap<uint32_t, Value*>& d) {
   errs() << "{\n";
   for (DenseMap<uint32_t, Value*>::iterator I = d.begin(),
        E = d.end(); I != E; ++I) {
       errs() << I->first << "\n";