[llvm] r279829 - [LoopUnroll] Use OptimizationRemarkEmitter directly not via the analysis pass

[Adam Nemet via llvm-commits]

Author: anemet
Date: Fri Aug 26 10:58:34 2016
New Revision: 279829

URL: http://llvm.org/viewvc/llvm-project?rev=279829&view=rev
Log:
[LoopUnroll] Use OptimizationRemarkEmitter directly not via the analysis pass

We can't mark ORE (a function pass) preserved as required by the loop
passes because that is how we ensure that the required passes like
LazyBFI are all available any time ORE is used.  See the new comments in
the patch.

Instead we use it directly just like the inliner does in D22694.

As expected there is some additional overhead after removing the caching
provided by analysis passes.  The worst case, I measured was
LNT/CINT2006_ref/401.bzip2 which regresses by 12%.  As before, this only
affects -Rpass-with-hotness and not default compilation.

Modified:
    llvm/trunk/include/llvm/Analysis/OptimizationDiagnosticInfo.h
    llvm/trunk/lib/Transforms/Scalar/LoopUnrollPass.cpp
    llvm/trunk/lib/Transforms/Utils/LoopUtils.cpp

Modified: llvm/trunk/include/llvm/Analysis/OptimizationDiagnosticInfo.h
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/Analysis/OptimizationDiagnosticInfo.h?rev=279829&r1=279828&r2=279829&view=diff
==============================================================================
--- llvm/trunk/include/llvm/Analysis/OptimizationDiagnosticInfo.h (original)
+++ llvm/trunk/include/llvm/Analysis/OptimizationDiagnosticInfo.h Fri Aug 26 10:58:34 2016
@@ -177,6 +177,11 @@ private:
   void operator=(const OptimizationRemarkEmitter &) = delete;
 };
 
+/// OptimizationRemarkEmitter legacy analysis pass
+///
+/// Note that this pass shouldn't generally be marked as preserved by other
+/// passes.  It's holding onto BFI, so if the pass does not preserve BFI, BFI
+/// could be freed.
 class OptimizationRemarkEmitterWrapperPass : public FunctionPass {
   std::unique_ptr<OptimizationRemarkEmitter> ORE;
 

Modified: llvm/trunk/lib/Transforms/Scalar/LoopUnrollPass.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/LoopUnrollPass.cpp?rev=279829&r1=279828&r2=279829&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/Scalar/LoopUnrollPass.cpp (original)
+++ llvm/trunk/lib/Transforms/Scalar/LoopUnrollPass.cpp Fri Aug 26 10:58:34 2016
@@ -1018,7 +1018,10 @@ public:
     const TargetTransformInfo &TTI =
         getAnalysis<TargetTransformInfoWrapperPass>().getTTI(F);
     auto &AC = getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F);
-    auto &ORE = getAnalysis<OptimizationRemarkEmitterWrapperPass>().getORE();
+    // For the old PM, we can't use OptimizationRemarkEmitter as an analysis
+    // pass.  Function analyses need to be preserved across loop transformations
+    // but ORE cannot be preserved (see comment before the pass definition).
+    OptimizationRemarkEmitter ORE(&F);
     bool PreserveLCSSA = mustPreserveAnalysisID(LCSSAID);
 
     return tryToUnrollLoop(L, DT, LI, SE, TTI, AC, ORE, PreserveLCSSA,

Modified: llvm/trunk/lib/Transforms/Utils/LoopUtils.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Utils/LoopUtils.cpp?rev=279829&r1=279828&r2=279829&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/Utils/LoopUtils.cpp (original)
+++ llvm/trunk/lib/Transforms/Utils/LoopUtils.cpp Fri Aug 26 10:58:34 2016
@@ -17,7 +17,6 @@
 #include "llvm/Analysis/GlobalsModRef.h"
 #include "llvm/Analysis/GlobalsModRef.h"
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/OptimizationDiagnosticInfo.h"
 #include "llvm/Analysis/ScalarEvolution.h"
 #include "llvm/Analysis/ScalarEvolutionAliasAnalysis.h"
 #include "llvm/Analysis/ScalarEvolutionExpander.h"
@@ -962,8 +961,6 @@ void llvm::getLoopAnalysisUsage(Analysis
   AU.addPreserved<SCEVAAWrapperPass>();
   AU.addRequired<ScalarEvolutionWrapperPass>();
   AU.addPreserved<ScalarEvolutionWrapperPass>();
-  AU.addRequired<OptimizationRemarkEmitterWrapperPass>();
-  AU.addPreserved<OptimizationRemarkEmitterWrapperPass>();
 }
 
 /// Manually defined generic "LoopPass" dependency initialization. This is used
@@ -984,7 +981,6 @@ void llvm::initializeLoopPassPass(PassRe
   INITIALIZE_PASS_DEPENDENCY(GlobalsAAWrapperPass)
   INITIALIZE_PASS_DEPENDENCY(SCEVAAWrapperPass)
   INITIALIZE_PASS_DEPENDENCY(ScalarEvolutionWrapperPass)
-  INITIALIZE_PASS_DEPENDENCY(OptimizationRemarkEmitterWrapperPass)
 }
 
 /// \brief Find string metadata for loop