New pass manager for LICM.

[danielcdh]

Summary

Port LICM to the new pass manager.

[davide]

The new LoopInvariantCodeMotion struct/class needs to live in an header in the new PM.
You also don't need to convert all the tests, just one or two of them should be enough.

[danielcdh]

include the header file.

[davidxl]

lib/Transforms/Scalar/LICM.cpp

139

getAnalysisIfAvailable(..)

149

This changes functionality -- remove

177

add a //FIXME for skip loop check

287–288

guarded with null check?

test/Transforms/LICM/hoist-deref-load.ll

2

Is loop-simplifycfg available in new pm?

[davide]

Some comments.

include/llvm/Transforms/Scalar/LICM.h

2

LoopSimplifyCFG.cpp -> LICM.h ?

42

This comment is wrong, no?

lib/Transforms/Scalar/LICM.cpp

154

There's no doFinalization() helper in the new PM (but the logic should be either removed or moved to the destructor). Can you make sure this happen? I think this check is kinda important here.

186

I assume TLIi is TLI? Does this code even build?

192

The old PM preserves the CFG add a fix for that please.

[silvas]

lib/Transforms/Scalar/LICM.cpp

1130–1132

This seems to be called by the LPM in the old pass manager as a notification. I don't think there is an analog in the new PM -- does there need to be? What happens if this is not called?

[danielcdh]

integrate comments.

lib/Transforms/Scalar/LICM.cpp

154

The doFinalization performs a legality check for the caching data structure LoopToAliasSetMap. However, as the new pass manage is lack of equivalence of the following helper functions:

cloneBasicBlockAnalysis
deleteAnalysisValue
deleteAnalysisLoop

As a result, LoopToAliasSetMap was not used in the new pass manager implementation.

I've added a FIXME to collectAliasInfoForLoop to clarify this and remind hooking it up once available.

192

Added a FIXME.

1130–1132

For LICM, these helper functions are for compile time purposes: avoid recompute the AST for the inner loop when processing the outer loop. And you are right, these notifications does not exist in new PM. It would be best if we can have that in new PM. But without it, when processing the outer loop, it will just recompute the AST for the inner loop. So it will affect compile time but not generated code.

test/Transforms/LICM/hoist-deref-load.ll

2

Yes, it is. But I'm not sure how to specify to make it run prior to licm in the -passes= option. any ideas?

[silvas]

lib/Transforms/Scalar/LICM.cpp

1130–1133

Thanks for taking a look at this. If it is just a compile time optimization then it is not a big deal (unless it is a huge difference). We can introduce a similar mechanism in the new PM if this shows up in the profile.

test/Transforms/LICM/hoist-deref-load.ll

2

You can do loop(simplify-cfg,licm)

[davidxl]

lgtm with the test fix suggested by sean: loop(simplify-cfg,licm)

[danielcdh]

oops, drafted the comments few days ago but forgot to hit "Submit"...

test/Transforms/LICM/hoist-deref-load.ll

2

Unfortunately, the loop-simplifycfg pass behaviors quite differently under the new PM:

opt -passes='require<domtree>,loop(simplify-cfg)' test/Transforms/LICM/hoist-deref-load.ll -S > 1
opt -loop-simplifycfg test/Transforms/LICM/hoist-deref-load.ll -S > 2

The output 1 and 2 are quite different. Am I missing something when invoking it in the new PM?

[davidxl]

test/Transforms/LICM/hoist-deref-load.ll

2

it also depends on loop analysis and loop canonicalization (loop simplification). The later is not ported yet.

So for now just add a FIXME there and proceed with your original change.

[silvas]

test/Transforms/LICM/hoist-deref-load.ll

2

Davide recently committed LoopSimplify in http://reviews.llvm.org/D21404 / r274959. Dehao, you should be able to add an explicit call to LoopSimplify in order to get the result you want.

[danielcdh]

migrate loop-simplifycfg pass to new pass manager in the licm tests.

[silvas]

LGTM

[vitalybuka]

This patch adds memory leaks:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/14569/steps/check-llvm%20asan/logs/stdio
Pleas let me know if you have a fix or I can revert it.

[krasin]

FYI: this seems to have broken sanitizers bot: http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/14569

==13022==ERROR: LeakSanitizer: detected memory leaks

Direct leak of 48 byte(s) in 1 object(s) allocated from:
    #0 0x84b1b0 in operator new(unsigned long) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/projects/compiler-rt/lib/asan/asan_new_delete.cc:78
    #1 0x38a8df3 in collectAliasInfoForLoop /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:1105:14
    #2 0x38a8df3 in (anonymous namespace)::LoopInvariantCodeMotion::runOnLoop(llvm::Loop*, llvm::AAResults*, llvm::LoopInfo*, llvm::DominatorTree*, llvm::TargetLibraryInfo*, llvm::ScalarEvolution*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:222
    #3 0x38a818d in llvm::LICMPass::run(llvm::Loop&, llvm::AnalysisManager<llvm::Loop>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:192:13
    #4 0x40f93f1 in llvm::detail::PassModel<llvm::Loop, llvm::LICMPass, llvm::PreservedAnalyses>::run(llvm::Loop&, llvm::AnalysisManager<llvm::Loop>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManagerInternal.h:72:17
    #5 0x4c78ee5 in llvm::PassManager<llvm::Loop>::run(llvm::Loop&, llvm::AnalysisManager<llvm::Loop>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:255:47
    #6 0x410108a in llvm::FunctionToLoopPassAdaptor<llvm::PassManager<llvm::Loop> >::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/Analysis/LoopPassManager.h:106:39
    #7 0x4100611 in llvm::detail::PassModel<llvm::Function, llvm::FunctionToLoopPassAdaptor<llvm::PassManager<llvm::Loop> >, llvm::PreservedAnalyses>::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManagerInternal.h:72:17
    #8 0x316c921 in llvm::PassManager<llvm::Function>::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:255:47
    #9 0x4092db5 in llvm::ModuleToFunctionPassAdaptor<llvm::PassManager<llvm::Function> >::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:900:39
    #10 0x4092a01 in llvm::detail::PassModel<llvm::Module, llvm::ModuleToFunctionPassAdaptor<llvm::PassManager<llvm::Function> >, llvm::PreservedAnalyses>::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManagerInternal.h:72:17
    #11 0x316aff8 in llvm::PassManager<llvm::Module>::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:255:47
    #12 0x85311c in llvm::runPassPipeline(llvm::StringRef, llvm::LLVMContext&, llvm::Module&, llvm::TargetMachine*, llvm::tool_output_file*, llvm::StringRef, llvm::opt_tool::OutputKind, llvm::opt_tool::VerifierKind, bool, bool) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/opt/NewPMDriver.cpp:112:7
    #13 0x87e701 in main /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/opt/opt.cpp:469:12
    #14 0x7effe31eaf44 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x21f44)

Indirect leak of 3072 byte(s) in 1 object(s) allocated from:
    #0 0x84b1b0 in operator new(unsigned long) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/projects/compiler-rt/lib/asan/asan_new_delete.cc:78
    #1 0x2247e5a in allocateBuckets /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/DenseMap.h:731:37
    #2 0x2247e5a in llvm::DenseMap<llvm::AliasSetTracker::ASTCallbackVH, llvm::AliasSet::PointerRec*, llvm::AliasSetTracker::ASTCallbackVHDenseMapInfo, llvm::detail::DenseMapPair<llvm::AliasSetTracker::ASTCallbackVH, llvm::AliasSet::PointerRec*> >::grow(unsigned int) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/DenseMap.h:671
    #3 0x2247c51 in grow /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/DenseMap.h:425:36
    #4 0x2247c51 in llvm::detail::DenseMapPair<llvm::AliasSetTracker::ASTCallbackVH, llvm::AliasSet::PointerRec*>* llvm::DenseMapBase<llvm::DenseMap<llvm::AliasSetTracker::ASTCallbackVH, llvm::AliasSet::PointerRec*, llvm::AliasSetTracker::ASTCallbackVHDenseMapInfo, llvm::detail::DenseMapPair<llvm::AliasSetTracker::ASTCallbackVH, llvm::AliasSet::PointerRec*> >, llvm::AliasSetTracker::ASTCallbackVH, llvm::AliasSet::PointerRec*, llvm::AliasSetTracker::ASTCallbackVHDenseMapInfo, llvm::detail::DenseMapPair<llvm::AliasSetTracker::ASTCallbackVH, llvm::AliasSet::PointerRec*> >::InsertIntoBucketImpl<llvm::AliasSetTracker::ASTCallbackVH>(llvm::AliasSetTracker::ASTCallbackVH const&, llvm::AliasSetTracker::ASTCallbackVH const&, llvm::detail::DenseMapPair<llvm::AliasSetTracker::ASTCallbackVH, llvm::AliasSet::PointerRec*>*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/DenseMap.h:486
    #5 0x22408cc in InsertIntoBucket /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/DenseMap.h:452:17
    #6 0x22408cc in FindAndConstruct /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/DenseMap.h:264
    #7 0x22408cc in operator[] /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/DenseMap.h:268
    #8 0x22408cc in llvm::AliasSetTracker::getEntryFor(llvm::Value*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/Analysis/AliasSetTracker.h:420
    #9 0x223fff6 in llvm::AliasSetTracker::getAliasSetForPointer(llvm::Value*, unsigned long, llvm::AAMDNodes const&, bool*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Analysis/AliasSetTracker.cpp:272:33
    #10 0x2241601 in addPointer /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/Analysis/AliasSetTracker.h:430:20
    #11 0x2241601 in llvm::AliasSetTracker::add(llvm::StoreInst*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Analysis/AliasSetTracker.cpp:333
    #12 0x2241f61 in llvm::AliasSetTracker::add(llvm::Instruction*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Analysis/AliasSetTracker.cpp:391:12
    #13 0x22421c2 in llvm::AliasSetTracker::add(llvm::BasicBlock&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Analysis/AliasSetTracker.cpp:402:5
    #14 0x38a917e in operator() /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:1113:17
    #15 0x38a917e in collectAliasInfoForLoop /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:1121
    #16 0x38a917e in (anonymous namespace)::LoopInvariantCodeMotion::runOnLoop(llvm::Loop*, llvm::AAResults*, llvm::LoopInfo*, llvm::DominatorTree*, llvm::TargetLibraryInfo*, llvm::ScalarEvolution*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:222
    #17 0x38a818d in llvm::LICMPass::run(llvm::Loop&, llvm::AnalysisManager<llvm::Loop>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:192:13
    #18 0x40f93f1 in llvm::detail::PassModel<llvm::Loop, llvm::LICMPass, llvm::PreservedAnalyses>::run(llvm::Loop&, llvm::AnalysisManager<llvm::Loop>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManagerInternal.h:72:17
    #19 0x4c78ee5 in llvm::PassManager<llvm::Loop>::run(llvm::Loop&, llvm::AnalysisManager<llvm::Loop>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:255:47
    #20 0x410108a in llvm::FunctionToLoopPassAdaptor<llvm::PassManager<llvm::Loop> >::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/Analysis/LoopPassManager.h:106:39
    #21 0x4100611 in llvm::detail::PassModel<llvm::Function, llvm::FunctionToLoopPassAdaptor<llvm::PassManager<llvm::Loop> >, llvm::PreservedAnalyses>::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManagerInternal.h:72:17
    #22 0x316c921 in llvm::PassManager<llvm::Function>::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:255:47
    #23 0x4092db5 in llvm::ModuleToFunctionPassAdaptor<llvm::PassManager<llvm::Function> >::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:900:39
    #24 0x4092a01 in llvm::detail::PassModel<llvm::Module, llvm::ModuleToFunctionPassAdaptor<llvm::PassManager<llvm::Function> >, llvm::PreservedAnalyses>::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManagerInternal.h:72:17
    #25 0x316aff8 in llvm::PassManager<llvm::Module>::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:255:47
    #26 0x85311c in llvm::runPassPipeline(llvm::StringRef, llvm::LLVMContext&, llvm::Module&, llvm::TargetMachine*, llvm::tool_output_file*, llvm::StringRef, llvm::opt_tool::OutputKind, llvm::opt_tool::VerifierKind, bool, bool) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/opt/NewPMDriver.cpp:112:7
    #27 0x87e701 in main /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/opt/opt.cpp:469:12
    #28 0x7effe31eaf44 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x21f44)

Indirect leak of 72 byte(s) in 1 object(s) allocated from:
    #0 0x84b1b0 in operator new(unsigned long) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/projects/compiler-rt/lib/asan/asan_new_delete.cc:78
    #1 0x223f6e4 in createSentinel /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/ilist.h:78:44
    #2 0x223f6e4 in ensureHead /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/ilist.h:93
    #3 0x223f6e4 in CreateLazySentinel /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/ilist.h:324
    #4 0x223f6e4 in begin /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/ADT/ilist.h:357
    #5 0x223f6e4 in begin /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/Analysis/AliasSetTracker.h:407
    #6 0x223f6e4 in llvm::AliasSetTracker::mergeAliasSetsForPointer(llvm::Value const*, unsigned long, llvm::AAMDNodes const&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Analysis/AliasSetTracker.cpp:218
    #7 0x224006c in llvm::AliasSetTracker::getAliasSetForPointer(llvm::Value*, unsigned long, llvm::AAMDNodes const&, bool*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Analysis/AliasSetTracker.cpp:287:22
    #8 0x2241601 in addPointer /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/Analysis/AliasSetTracker.h:430:20
    #9 0x2241601 in llvm::AliasSetTracker::add(llvm::StoreInst*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Analysis/AliasSetTracker.cpp:333
    #10 0x2241f61 in llvm::AliasSetTracker::add(llvm::Instruction*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Analysis/AliasSetTracker.cpp:391:12
    #11 0x22421c2 in llvm::AliasSetTracker::add(llvm::BasicBlock&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Analysis/AliasSetTracker.cpp:402:5
    #12 0x38a917e in operator() /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:1113:17
    #13 0x38a917e in collectAliasInfoForLoop /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:1121
    #14 0x38a917e in (anonymous namespace)::LoopInvariantCodeMotion::runOnLoop(llvm::Loop*, llvm::AAResults*, llvm::LoopInfo*, llvm::DominatorTree*, llvm::TargetLibraryInfo*, llvm::ScalarEvolution*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:222
    #15 0x38a818d in llvm::LICMPass::run(llvm::Loop&, llvm::AnalysisManager<llvm::Loop>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Transforms/Scalar/LICM.cpp:192:13
    #16 0x40f93f1 in llvm::detail::PassModel<llvm::Loop, llvm::LICMPass, llvm::PreservedAnalyses>::run(llvm::Loop&, llvm::AnalysisManager<llvm::Loop>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManagerInternal.h:72:17
    #17 0x4c78ee5 in llvm::PassManager<llvm::Loop>::run(llvm::Loop&, llvm::AnalysisManager<llvm::Loop>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:255:47
    #18 0x410108a in llvm::FunctionToLoopPassAdaptor<llvm::PassManager<llvm::Loop> >::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/Analysis/LoopPassManager.h:106:39
    #19 0x4100611 in llvm::detail::PassModel<llvm::Function, llvm::FunctionToLoopPassAdaptor<llvm::PassManager<llvm::Loop> >, llvm::PreservedAnalyses>::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManagerInternal.h:72:17
    #20 0x316c921 in llvm::PassManager<llvm::Function>::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:255:47
    #21 0x4092db5 in llvm::ModuleToFunctionPassAdaptor<llvm::PassManager<llvm::Function> >::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:900:39
    #22 0x4092a01 in llvm::detail::PassModel<llvm::Module, llvm::ModuleToFunctionPassAdaptor<llvm::PassManager<llvm::Function> >, llvm::PreservedAnalyses>::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManagerInternal.h:72:17
    #23 0x316aff8 in llvm::PassManager<llvm::Module>::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/include/llvm/IR/PassManager.h:255:47
    #24 0x85311c in llvm::runPassPipeline(llvm::StringRef, llvm::LLVMContext&, llvm::Module&, llvm::TargetMachine*, llvm::tool_output_file*, llvm::StringRef, llvm::opt_tool::OutputKind, llvm::opt_tool::VerifierKind, bool, bool) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/opt/NewPMDriver.cpp:112:7
    #25 0x87e701 in main /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/opt/opt.cpp:469:12
    #26 0x7effe31eaf44 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x21f44)

SUMMARY: AddressSanitizer: 3192 byte(s) leaked in 3 allocation(s).

[vitalybuka]

Reverted http://reviews.llvm.org/D22259

[danielcdh]

Fix the memory leak.

[silvas]

LGTM. Give it another shot, but keep an eye on those bots that failed last time.

[danielcdh]

Tried to build clang with -DLLVM_USE_SANITIZER=Address, and verified the failing tests no longer fail with the updated patch.

include/llvm/InitializePasses.h

 void initializeInstSimplifierPass(PassRegistry&);
 void initializeInstrProfilingLegacyPassPass(PassRegistry &);
 void initializeInstructionCombiningPassPass(PassRegistry&);
 void initializeInterleavedAccessPass(PassRegistry &);
 void initializeInternalizeLegacyPassPass(PassRegistry&);
 void initializeIntervalPartitionPass(PassRegistry&);
 void initializeJumpThreadingPass(PassRegistry&);
 void initializeLCSSAWrapperPassPass(PassRegistry &);
-void initializeLICMPass(PassRegistry&);
+void initializeLegacyLICMPassPass(PassRegistry&);
 void initializeLazyValueInfoWrapperPassPass(PassRegistry&);
 void initializeLintPass(PassRegistry&);
 void initializeLiveDebugValuesPass(PassRegistry&);
 void initializeLiveDebugVariablesPass(PassRegistry&);
 void initializeLiveIntervalsPass(PassRegistry&);
 void initializeLiveRegMatrixPass(PassRegistry&);
 void initializeLiveStacksPass(PassRegistry&);
 void initializeLiveVariablesPass(PassRegistry&);

include/llvm/Transforms/Scalar/LICM.h

+//===- LICM.h - Loop Invariant Code Motion Pass -------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This pass performs loop invariant code motion, attempting to remove as much
+// code from the body of a loop as possible.  It does this by either hoisting
+// code into the preheader block, or by sinking code to the exit blocks if it is
+// safe.  This pass also promotes must-aliased memory locations in the loop to
+// live in registers, thus hoisting and sinking "invariant" loads and stores.
+//
+// This pass uses alias analysis for two purposes:
+//
+//  1. Moving loop invariant loads and calls out of loops.  If we can determine
+//     that a load or call inside of a loop never aliases anything stored to,
+//     we can hoist it or sink it like any other instruction.
+//  2. Scalar Promotion of Memory - If there is a store instruction inside of
+//     the loop, we try to move the store to happen AFTER the loop instead of
+//     inside of the loop.  This can only happen if a few conditions are true:
+//       A. The pointer stored through is loop invariant
+//       B. There are no stores or loads in the loop which _may_ alias the
+//          pointer.  There are no calls in the loop which mod/ref the pointer.
+//     If these conditions are true, we can promote the loads and stores in the
+//     loop of the pointer to use a temporary alloca'd variable.  We then use
+//     the SSAUpdater to construct the appropriate SSA form for the value.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_TRANSFORMS_SCALAR_LICM_H
+#define LLVM_TRANSFORMS_SCALAR_LICM_H
+
+#include "llvm/Analysis/LoopInfo.h"
+#include "llvm/IR/PassManager.h"
+
+namespace llvm {
+
+/// Performs Loop Invariant Code Motion Pass.
+class LICMPass : public PassInfoMixin<LICMPass> {
+public:
+  PreservedAnalyses run(Loop &L, AnalysisManager<Loop> &AM);
+};
+} // end namespace llvm
+
+#endif // LLVM_TRANSFORMS_SCALAR_LICM_H

lib/LTO/LTOCodeGenerator.cpp

void LTOCodeGenerator::initializeLTOPasses() {
   initializePruneEHPass(R);
   initializeGlobalDCELegacyPassPass(R);
   initializeArgPromotionPass(R);
   initializeJumpThreadingPass(R);
   initializeSROALegacyPassPass(R);
   initializePostOrderFunctionAttrsLegacyPassPass(R);
   initializeReversePostOrderFunctionAttrsLegacyPassPass(R);
   initializeGlobalsAAWrapperPassPass(R);
-  initializeLICMPass(R);
+  initializeLegacyLICMPassPass(R);
   initializeMergedLoadStoreMotionLegacyPassPass(R);
   initializeGVNLegacyPassPass(R);
   initializeMemCpyOptLegacyPassPass(R);
   initializeDCELegacyPassPass(R);
   initializeCFGSimplifyPassPass(R);
 }
 
 bool LTOCodeGenerator::addModule(LTOModule *Mod) {

lib/Transforms/Scalar/LICM.cpp

 //       B. There are no stores or loads in the loop which _may_ alias the
 //          pointer.  There are no calls in the loop which mod/ref the pointer.
 //     If these conditions are true, we can promote the loads and stores in the
 //     loop of the pointer to use a temporary alloca'd variable.  We then use
 //     the SSAUpdater to construct the appropriate SSA form for the value.
 //
 //===----------------------------------------------------------------------===//
 
+#include "llvm/Transforms/Scalar/LICM.h"
 #include "llvm/ADT/Statistic.h"
 #include "llvm/Analysis/AliasAnalysis.h"
 #include "llvm/Analysis/AliasSetTracker.h"
 #include "llvm/Analysis/BasicAliasAnalysis.h"
 #include "llvm/Analysis/CaptureTracking.h"
 #include "llvm/Analysis/ConstantFolding.h"
 #include "llvm/Analysis/GlobalsModRef.h"
 #include "llvm/Analysis/Loads.h"
 #include "llvm/Analysis/LoopInfo.h"
 #include "llvm/Analysis/LoopPass.h"
+#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/Analysis/MemoryBuiltins.h"
 #include "llvm/Analysis/ScalarEvolution.h"
 #include "llvm/Analysis/ScalarEvolutionAliasAnalysis.h"
 #include "llvm/Analysis/TargetLibraryInfo.h"
 #include "llvm/Analysis/ValueTracking.h"
 #include "llvm/IR/CFG.h"
 #include "llvm/IR/Constants.h"
 #include "llvm/IR/DataLayout.h"
CloneInstructionInExitBlock(Instruction &I, BasicBlock &ExitBlock, PHINode &PN,
                             const LoopInfo *LI,
                             const LoopSafetyInfo *SafetyInfo);
 static bool canSinkOrHoistInst(Instruction &I, AliasAnalysis *AA,
                                DominatorTree *DT, TargetLibraryInfo *TLI,
                                Loop *CurLoop, AliasSetTracker *CurAST,
                                LoopSafetyInfo *SafetyInfo);
 
 namespace {
-struct LICM : public LoopPass {
+struct LoopInvariantCodeMotion {
+  bool runOnLoop(Loop *L, AliasAnalysis *AA, LoopInfo *LI, DominatorTree *DT,
+                 TargetLibraryInfo *TLI, ScalarEvolution *SE, bool DeleteAST);
+
+  DenseMap<Loop *, AliasSetTracker *> &getLoopToAliasSetMap() {
+    return LoopToAliasSetMap;
+  }
+
+private:
+  DenseMap<Loop *, AliasSetTracker *> LoopToAliasSetMap;
+
+  AliasSetTracker *collectAliasInfoForLoop(Loop *L, LoopInfo *LI,
+                                           AliasAnalysis *AA);
+};
+
+struct LegacyLICMPass : public LoopPass {
   static char ID; // Pass identification, replacement for typeid
-  LICM() : LoopPass(ID) {
-    initializeLICMPass(*PassRegistry::getPassRegistry());
+  LegacyLICMPass() : LoopPass(ID) {
+    initializeLegacyLICMPassPass(*PassRegistry::getPassRegistry());
   }
 
-  bool runOnLoop(Loop *L, LPPassManager &LPM) override;
+  bool runOnLoop(Loop *L, LPPassManager &LPM) override {
+    if (skipLoop(L))
+      return false;
+
+    auto *SE = getAnalysisIfAvailable<ScalarEvolutionWrapperPass>();
+    return LICM.runOnLoop(L,
+                          &getAnalysis<AAResultsWrapperPass>().getAAResults(),
+                          &getAnalysis<LoopInfoWrapperPass>().getLoopInfo(),
+                          &getAnalysis<DominatorTreeWrapperPass>().getDomTree(),
+                          &getAnalysis<TargetLibraryInfoWrapperPass>().getTLI(),
+                          SE ? &SE->getSE() : nullptr, false);
+  }
 
   /// This transformation requires natural loop information & requires that
   /// loop preheaders be inserted into the CFG...
   ///
   void getAnalysisUsage(AnalysisUsage &AU) const override {
     AU.setPreservesCFG();
     AU.addRequired<TargetLibraryInfoWrapperPass>();
     getLoopAnalysisUsage(AU);
   }
 
   using llvm::Pass::doFinalization;
 
   bool doFinalization() override {
-    assert(LoopToAliasSetMap.empty() && "Didn't free loop alias sets");
+    assert(LICM.getLoopToAliasSetMap().empty() &&
+           "Didn't free loop alias sets");
     return false;
   }
 
 private:
-  AliasAnalysis *AA; // Current AliasAnalysis information
-  LoopInfo *LI;      // Current LoopInfo
-  DominatorTree *DT; // Dominator Tree for the current Loop.
-
-  TargetLibraryInfo *TLI; // TargetLibraryInfo for constant folding.
-
-  // State that is updated as we process loops.
-  bool Changed;            // Set to true when we change anything.
-  BasicBlock *Preheader;   // The preheader block of the current loop...
-  Loop *CurLoop;           // The current loop we are working on...
-  AliasSetTracker *CurAST; // AliasSet information for the current loop...
-  DenseMap<Loop *, AliasSetTracker *> LoopToAliasSetMap;
+  LoopInvariantCodeMotion LICM;
 
   /// cloneBasicBlockAnalysis - Simple Analysis hook. Clone alias set info.
   void cloneBasicBlockAnalysis(BasicBlock *From, BasicBlock *To,
                                Loop *L) override;
 
   /// deleteAnalysisValue - Simple Analysis hook. Delete value V from alias
   /// set.
   void deleteAnalysisValue(Value *V, Loop *L) override;
 
   /// Simple Analysis hook. Delete loop L from alias set map.
   void deleteAnalysisLoop(Loop *L) override;
-
-  AliasSetTracker *collectAliasInfoForLoop(Loop *L);
 };
 }
 
-char LICM::ID = 0;
-INITIALIZE_PASS_BEGIN(LICM, "licm", "Loop Invariant Code Motion", false, false)
+PreservedAnalyses LICMPass::run(Loop &L, AnalysisManager<Loop> &AM) {
+  // FIXME: Check if loop should be skipped.
+
+  const auto &FAM =
+      AM.getResult<FunctionAnalysisManagerLoopProxy>(L).getManager();
+  Function *F = L.getHeader()->getParent();
+
+  auto *AA = FAM.getCachedResult<AAManager>(*F);
+  auto *LI = FAM.getCachedResult<LoopAnalysis>(*F);
+  auto *DT = FAM.getCachedResult<DominatorTreeAnalysis>(*F);
+  auto *TLI = FAM.getCachedResult<TargetLibraryAnalysis>(*F);
+  auto *SE = FAM.getCachedResult<ScalarEvolutionAnalysis>(*F);
+  assert((AA && LI && DT && TLI && SE) && "Analyses for LICM not available");
+
+  LoopInvariantCodeMotion LICM;
+
+  if (!LICM.runOnLoop(&L, AA, LI, DT, TLI, SE, true))
+    return PreservedAnalyses::all();
+
+  // FIXME: There is no setPreservesCFG in the new PM. When that becomes
+  // available, it should be used here.
+  return getLoopPassPreservedAnalyses();
+}
+
+char LegacyLICMPass::ID = 0;
+INITIALIZE_PASS_BEGIN(LegacyLICMPass, "licm", "Loop Invariant Code Motion",
+                      false, false)
 INITIALIZE_PASS_DEPENDENCY(LoopPass)
 INITIALIZE_PASS_DEPENDENCY(TargetLibraryInfoWrapperPass)
-INITIALIZE_PASS_END(LICM, "licm", "Loop Invariant Code Motion", false, false)
+INITIALIZE_PASS_END(LegacyLICMPass, "licm", "Loop Invariant Code Motion", false,
+                    false)
 
-Pass *llvm::createLICMPass() { return new LICM(); }
+Pass *llvm::createLICMPass() { return new LegacyLICMPass(); }
 
 /// Hoist expressions out of the specified loop. Note, alias info for inner
 /// loop is not preserved so it is not a good idea to run LICM multiple
 /// times on one loop.
+/// We should delete AST for inner loops in the new pass manager to avoid
+/// memory leak.
 ///
-bool LICM::runOnLoop(Loop *L, LPPassManager &LPM) {
-  if (skipLoop(L))
-    return false;
-
-  Changed = false;
-
-  // Get our Loop and Alias Analysis information...
-  LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
-  AA = &getAnalysis<AAResultsWrapperPass>().getAAResults();
-  DT = &getAnalysis<DominatorTreeWrapperPass>().getDomTree();
-
-  TLI = &getAnalysis<TargetLibraryInfoWrapperPass>().getTLI();
+bool LoopInvariantCodeMotion::runOnLoop(Loop *L, AliasAnalysis *AA,
+                                        LoopInfo *LI, DominatorTree *DT,
+                                        TargetLibraryInfo *TLI,
+                                        ScalarEvolution *SE, bool DeleteAST) {
+  bool Changed = false;
 
   assert(L->isLCSSAForm(*DT) && "Loop is not in LCSSA form.");
 
-  CurAST = collectAliasInfoForLoop(L);
-
-  CurLoop = L;
+  AliasSetTracker *CurAST = collectAliasInfoForLoop(L, LI, AA);
 
   // Get the preheader block to move instructions into...
-  Preheader = L->getLoopPreheader();
+  BasicBlock *Preheader = L->getLoopPreheader();
 
   // Compute loop safety information.
   LoopSafetyInfo SafetyInfo;
-  computeLoopSafetyInfo(&SafetyInfo, CurLoop);
+  computeLoopSafetyInfo(&SafetyInfo, L);
 
   // We want to visit all of the instructions in this loop... that are not parts
   // of our subloops (they have already had their invariants hoisted out of
   // their loop, into this loop, so there is no need to process the BODIES of
   // the subloops).
   //
   // Traverse the body of the loop in depth first order on the dominator tree so
   // that we are guaranteed to see definitions before we see uses.  This allows
   // us to sink instructions in one pass, without iteration.  After sinking
   // instructions, we perform another pass to hoist them out of the loop.
   //
   if (L->hasDedicatedExits())
-    Changed |= sinkRegion(DT->getNode(L->getHeader()), AA, LI, DT, TLI, CurLoop,
+    Changed |= sinkRegion(DT->getNode(L->getHeader()), AA, LI, DT, TLI, L,
                           CurAST, &SafetyInfo);
   if (Preheader)
-    Changed |= hoistRegion(DT->getNode(L->getHeader()), AA, LI, DT, TLI,
-                           CurLoop, CurAST, &SafetyInfo);
+    Changed |= hoistRegion(DT->getNode(L->getHeader()), AA, LI, DT, TLI, L,
+                           CurAST, &SafetyInfo);
 
   // Now that all loop invariants have been removed from the loop, promote any
   // memory references to scalars that we can.
   if (!DisablePromotion && (Preheader || L->hasDedicatedExits())) {
     SmallVector<BasicBlock *, 8> ExitBlocks;
     SmallVector<Instruction *, 8> InsertPts;
     PredIteratorCache PIC;
 
     // Loop over all of the alias sets in the tracker object.
     for (AliasSet &AS : *CurAST)
-      Changed |=
-          promoteLoopAccessesToScalars(AS, ExitBlocks, InsertPts, PIC, LI, DT,
-                                       TLI, CurLoop, CurAST, &SafetyInfo);
+      Changed |= promoteLoopAccessesToScalars(
+          AS, ExitBlocks, InsertPts, PIC, LI, DT, TLI, L, CurAST, &SafetyInfo);
 
     // Once we have promoted values across the loop body we have to recursively
     // reform LCSSA as any nested loop may now have values defined within the
     // loop used in the outer loop.
     // FIXME: This is really heavy handed. It would be a bit better to use an
     // SSAUpdater strategy during promotion that was LCSSA aware and reformed
     // it as it went.
     if (Changed) {
-      auto *SEWP = getAnalysisIfAvailable<ScalarEvolutionWrapperPass>();
-      formLCSSARecursively(*L, *DT, LI, SEWP ? &SEWP->getSE() : nullptr);
+      formLCSSARecursively(*L, *DT, LI, SE);
     }
   }
 
   // Check that neither this loop nor its parent have had LCSSA broken. LICM is
   // specifically moving instructions across the loop boundary and so it is
   // especially in need of sanity checking here.
   assert(L->isLCSSAForm(*DT) && "Loop not left in LCSSA form after LICM!");
   assert((!L->getParentLoop() || L->getParentLoop()->isLCSSAForm(*DT)) &&
          "Parent loop not left in LCSSA form after LICM!");
 
-  // Clear out loops state information for the next iteration
-  CurLoop = nullptr;
-  Preheader = nullptr;
-
   // If this loop is nested inside of another one, save the alias information
   // for when we process the outer loop.
-  if (L->getParentLoop())
+  if (L->getParentLoop() && !DeleteAST)
     LoopToAliasSetMap[L] = CurAST;
   else
     delete CurAST;
 
-  if (Changed)
-    if (auto *SEWP = getAnalysisIfAvailable<ScalarEvolutionWrapperPass>())
-      SEWP->getSE().forgetLoopDispositions(L);
+  if (Changed && SE)
+    SE->forgetLoopDispositions(L);
   return Changed;
 }
 
 /// Walk the specified region of the CFG (defined by all blocks dominated by
 /// the specified block, and that are in the current loop) in reverse depth
 /// first order w.r.t the DominatorTree.  This allows us to visit uses before
 /// definitions, allowing us to sink a loop body in one pass without iteration.
 ///
void llvm::computeLoopSafetyInfo(LoopSafetyInfo *SafetyInfo, Loop *CurLoop) {
   assert(CurLoop != nullptr && "CurLoop cant be null");
   BasicBlock *Header = CurLoop->getHeader();
   // Setting default safety values.
   SafetyInfo->MayThrow = false;
   SafetyInfo->HeaderMayThrow = false;
   // Iterate over header and compute safety info.
   for (BasicBlock::iterator I = Header->begin(), E = Header->end();
        (I != E) && !SafetyInfo->HeaderMayThrow; ++I)
-    SafetyInfo->HeaderMayThrow |= !isGuaranteedToTransferExecutionToSuccessor(&*I);
+    SafetyInfo->HeaderMayThrow |=
+        !isGuaranteedToTransferExecutionToSuccessor(&*I);
 
   SafetyInfo->MayThrow = SafetyInfo->HeaderMayThrow;
   // Iterate over loop instructions and compute safety info.
   for (Loop::block_iterator BB = CurLoop->block_begin(),
                             BBE = CurLoop->block_end();
        (BB != BBE) && !SafetyInfo->MayThrow; ++BB)
     for (BasicBlock::iterator I = (*BB)->begin(), E = (*BB)->end();
          (I != E) && !SafetyInfo->MayThrow; ++I)
bool llvm::promoteLoopAccessesToScalars(
   if (PreheaderLoad->use_empty())
     PreheaderLoad->eraseFromParent();
 
   return Changed;
 }
 
 /// Returns an owning pointer to an alias set which incorporates aliasing info
 /// from L and all subloops of L.
-AliasSetTracker *LICM::collectAliasInfoForLoop(Loop *L) {
+/// FIXME: In new pass manager, there is no helper functions to handle loop
+/// analysis such as cloneBasicBlockAnalysis. So the AST needs to be recompute
+/// from scratch for every loop. Hook up with the helper functions when
+/// available in the new pass manager to avoid redundant computation.
+AliasSetTracker *
+LoopInvariantCodeMotion::collectAliasInfoForLoop(Loop *L, LoopInfo *LI,
+                                                 AliasAnalysis *AA) {
   AliasSetTracker *CurAST = nullptr;
   SmallVector<Loop *, 4> RecomputeLoops;
   for (Loop *InnerL : L->getSubLoops()) {
     auto MapI = LoopToAliasSetMap.find(InnerL);
     // If the AST for this inner loop is missing it may have been merged into
     // some other loop's AST and then that loop unrolled, and so we need to
     // recompute it.
     if (MapI == LoopToAliasSetMap.end()) {
LoopInvariantCodeMotion::collectAliasInfoForLoop(Loop *L, LoopInfo *LI,
   // And merge in this loop.
   mergeLoop(L);
 
   return CurAST;
 }
 
 /// Simple analysis hook. Clone alias set info.
 ///
-void LICM::cloneBasicBlockAnalysis(BasicBlock *From, BasicBlock *To, Loop *L) {
-  AliasSetTracker *AST = LoopToAliasSetMap.lookup(L);
+void LegacyLICMPass::cloneBasicBlockAnalysis(BasicBlock *From, BasicBlock *To,
+                                             Loop *L) {
+  AliasSetTracker *AST = LICM.getLoopToAliasSetMap().lookup(L);
   if (!AST)
     return;
 
   AST->copyValue(From, To);
 }
 
 /// Simple Analysis hook. Delete value V from alias set
 ///
-void LICM::deleteAnalysisValue(Value *V, Loop *L) {
-  AliasSetTracker *AST = LoopToAliasSetMap.lookup(L);
+void LegacyLICMPass::deleteAnalysisValue(Value *V, Loop *L) {
+  AliasSetTracker *AST = LICM.getLoopToAliasSetMap().lookup(L);
   if (!AST)
     return;
 
   AST->deleteValue(V);
 }
 
 /// Simple Analysis hook. Delete value L from alias set map.
 ///
-void LICM::deleteAnalysisLoop(Loop *L) {
-  AliasSetTracker *AST = LoopToAliasSetMap.lookup(L);
+void LegacyLICMPass::deleteAnalysisLoop(Loop *L) {
+  AliasSetTracker *AST = LICM.getLoopToAliasSetMap().lookup(L);
   if (!AST)
     return;
 
   delete AST;
-  LoopToAliasSetMap.erase(L);
+  LICM.getLoopToAliasSetMap().erase(L);
 }
 
 /// Return true if the body of this loop may store into the memory
 /// location pointed to by V.
 ///
 static bool pointerInvalidatedByLoop(Value *V, uint64_t Size,
                                      const AAMDNodes &AAInfo,
                                      AliasSetTracker *CurAST) {

lib/Transforms/Scalar/Scalar.cpp

void llvm::initializeScalarOpts(PassRegistry &Registry) {
   initializeDSELegacyPassPass(Registry);
   initializeGuardWideningLegacyPassPass(Registry);
   initializeGVNLegacyPassPass(Registry);
   initializeEarlyCSELegacyPassPass(Registry);
   initializeFlattenCFGPassPass(Registry);
   initializeInductiveRangeCheckEliminationPass(Registry);
   initializeIndVarSimplifyLegacyPassPass(Registry);
   initializeJumpThreadingPass(Registry);
-  initializeLICMPass(Registry);
+  initializeLegacyLICMPassPass(Registry);
   initializeLoopDataPrefetchPass(Registry);
   initializeLoopDeletionPass(Registry);
   initializeLoopAccessLegacyAnalysisPass(Registry);
   initializeLoopInstSimplifyPass(Registry);
   initializeLoopInterchangePass(Registry);
   initializeLoopRotateLegacyPassPass(Registry);
   initializeLoopStrengthReducePass(Registry);
   initializeLoopRerollPass(Registry);

test/Transforms/LICM/argmemonly-call.ll

 ; RUN: opt -S -basicaa -licm %s | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' < %s -S | FileCheck %s
 declare i32 @foo() readonly argmemonly nounwind
 declare i32 @foo2() readonly nounwind
 declare i32 @bar(i32* %loc2) readonly argmemonly nounwind
 
 define void @test(i32* %loc) {
 ; CHECK-LABEL: @test
 ; CHECK: @foo
 ; CHECK-LABEL: loop:

test/Transforms/LICM/assume.ll

 ; RUN: opt -licm -basicaa < %s -S | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' < %s -S | FileCheck %s
 
 define void @f_0(i1 %p) nounwind ssp {
 ; CHECK-LABEL: @f_0(
 entry:
   br label %for.body
 
 for.body:
   br i1 undef, label %if.then, label %for.cond.backedge

test/Transforms/LICM/atomics.ll

 ; RUN: opt < %s -S -basicaa -licm | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='lcssa,require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' < %s -S | FileCheck %s
 
 ; Check that we can hoist unordered loads
 define i32 @test1(i32* nocapture %y) nounwind uwtable ssp {
 entry:
   br label %loop
 
 loop:
   %i = phi i32 [ %inc, %loop ], [ 0, %entry ]

test/Transforms/LICM/basictest.ll

 ; RUN: opt < %s -licm | llvm-dis
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' < %s | llvm-dis
 
 define void @testfunc(i32 %i) {
 ; <label>:0
   br label %Loop
 Loop:   ; preds = %Loop, %0
   %j = phi i32 [ 0, %0 ], [ %Next, %Loop ]    ; <i32> [#uses=1]
   %i2 = mul i32 %i, 17    ; <i32> [#uses=1]
   %Next = add i32 %j, %i2   ; <i32> [#uses=2]
   %cond = icmp eq i32 %Next, 0    ; <i1> [#uses=1]
   br i1 %cond, label %Out, label %Loop
 Out:    ; preds = %Loop
   ret void
 }
 

test/Transforms/LICM/constexpr.ll

 ; RUN: opt < %s -S -basicaa -licm | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='lcssa,require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' < %s -S | FileCheck %s
 ; This fixes PR22460
 
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-pc-windows-msvc"
 
 @in = internal unnamed_addr global i32* null, align 8
 @out = internal unnamed_addr global i32* null, align 8
 

test/Transforms/LICM/crash.ll

 ; RUN: opt -licm -disable-output < %s
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -disable-output < %s
 
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
 target triple = "x86_64-apple-darwin10.0.0"
 
 
 ; PR8068
 @g_12 = external global i8, align 1
 define void @test1() nounwind ssp {

test/Transforms/LICM/debug-value.ll

 ; RUN: opt -licm -basicaa < %s -S | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' < %s -S | FileCheck %s
 
 define void @dgefa() nounwind ssp {
 entry:
   br label %for.body
 
 for.body:                                         ; preds = %for.cond.backedge, %entry
   br i1 undef, label %if.then, label %for.cond.backedge, !dbg !11
 

test/Transforms/LICM/extra-copies.ll

 ; RUN: opt < %s -licm -S | FileCheck %s
+; RUN: opt -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' < %s -S | FileCheck %s
 ; PR19835
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-unknown-linux-gnu"
 
 define i32 @f(i32 %x) {
 entry:
   br label %for.body
 

test/Transforms/LICM/funclet.ll

 ; RUN: opt < %s -licm -S | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' < %s -S | FileCheck %s
 
 target datalayout = "e-m:x-p:32:32-i64:64-f80:32-n8:16:32-a:0:32-S32"
 target triple = "i386-pc-windows-msvc18.0.0"
 
 define void @test1(i32* %s, i1 %b) personality i32 (...)* @__CxxFrameHandler3 {
 entry:
   br label %while.cond
 

test/Transforms/LICM/hoist-bitcast-load.ll

 ; RUN: opt -S -basicaa -licm < %s | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='loop-simplify,require<aa>,require<targetir>,require<scalar-evolution>,loop(simplify-cfg,licm)' -S < %s | FileCheck %s
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-unknown-linux-gnu"
 
 ; Make sure the basic alloca pointer hoisting works:
 ; CHECK-LABEL: @test1
 ; CHECK: load i32, i32* %c, align 4
 ; CHECK: for.body:
 

test/Transforms/LICM/hoist-deref-load.ll

 ; RUN: opt -S -basicaa -licm < %s | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='loop-simplify,require<aa>,require<targetir>,require<scalar-evolution>,loop(simplify-cfg,licm)' -S < %s | FileCheck %s
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-unknown-linux-gnu"
 
 ; This test represents the following function:
 ; void test1(int * __restrict__ a, int * __restrict__ b, int &c, int n) {
 ;   for (int i = 0; i < n; ++i)
 ;     if (a[i] > 0)
 ;       a[i] = c*b[i];

test/Transforms/LICM/hoist-nounwind.ll

 ; RUN: opt -S -basicaa -licm < %s | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='lcssa,require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S %s | FileCheck %s
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-unknown-linux-gnu"
 
 declare void @f() nounwind
 
 ; Don't hoist load past nounwind call.
 define i32 @test1(i32* noalias nocapture readonly %a) nounwind uwtable {
 ; CHECK-LABEL: @test1(

test/Transforms/LICM/hoist-round.ll

 ; RUN: opt -S -licm < %s | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S %s | FileCheck %s
 
 target datalayout = "E-m:e-p:32:32-i8:8:8-i16:16:16-i64:32:32-f64:32:32-v64:32:32-v128:32:32-a0:0:32-n32"
 
 ; This test verifies that ceil, floor, nearbyint, trunc, rint, round,
 ; copysign, minnum, maxnum and fabs intrinsics are considered safe
 ; to speculate.
 
 ; CHECK-LABEL: @test

test/Transforms/LICM/hoisting.ll

 ; RUN: opt < %s -licm -S | FileCheck %s
+; RUN: opt -lcssa %s | opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S | FileCheck %s
 
 @X = global i32 0   ; <i32*> [#uses=1]
 
 declare void @foo()
 
 ; This testcase tests for a problem where LICM hoists 
 ; potentially trapping instructions when they are not guaranteed to execute.
 define i32 @test1(i1 %c) {

test/Transforms/LICM/lcssa-ssa-promoter.ll

 ; RUN: opt -S -basicaa -licm < %s | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S %s| FileCheck %s
 ;
 ; Manually validate LCSSA form is preserved even after SSAUpdater is used to
 ; promote things in the loop bodies.
 
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-unknown-linux-gnu"
 
 @x = common global i32 0, align 4

test/Transforms/LICM/no-preheader-test.ll

 ; Test that LICM works when there is not a loop-preheader
 ; RUN: opt < %s -licm | llvm-dis
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' < %s | llvm-dis
 
 define void @testfunc(i32 %i.s, i1 %ifcond) {
   br i1 %ifcond, label %Then, label %Else
 Then:   ; preds = %0
   br label %Loop
 Else:   ; preds = %0
   br label %Loop
 Loop:   ; preds = %Loop, %Else, %Then

test/Transforms/LICM/preheader-safe.ll

 ; RUN: opt -S -licm < %s | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S %s | FileCheck %s
 
 declare void @use_nothrow(i64 %a) nounwind
 declare void @use(i64 %a)
 
 define void @nothrow(i64 %x, i64 %y, i1* %cond) {
 ; CHECK-LABEL: nothrow
 ; CHECK-LABEL: entry
 ; CHECK: %div = udiv i64 %x, %y

test/Transforms/LICM/promote-order.ll

 ; RUN: opt -tbaa -basicaa -licm -S < %s | FileCheck %s
+; RUN: opt -aa-pipeline=type-based-aa,basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S %s | FileCheck %s
 
 ; LICM should keep the stores in their original order when it sinks/promotes them.
 ; rdar://12045203
 
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
 target triple = "x86_64-apple-macosx10.8.0"
 
 @p = external global i8*

test/Transforms/LICM/promote-tls.ll

 ; RUN: opt -tbaa -basicaa -licm -S < %s | FileCheck %s
+; RUN: opt -aa-pipeline=type-based-aa,basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S %s | FileCheck %s
 
 ; If we can prove a local is thread local, we can insert stores during
 ; promotion which wouldn't be legal otherwise.
 
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
 target triple = "x86_64-linux-generic"
 
 @p = external global i8*

test/Transforms/LICM/scalar-promote-memmodel.ll

 ; RUN: opt < %s -basicaa -licm -S | FileCheck %s
+; RUN: opt -aa-pipeline=type-based-aa,basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S %s | FileCheck %s
 
 ; Make sure we don't hoist a conditionally-executed store out of the loop;
 ; it would violate the concurrency memory model
 
 @g = common global i32 0, align 4
 
 define void @bar(i32 %n, i32 %b) nounwind uwtable ssp {
 entry:

test/Transforms/LICM/scalar_promote-unwind.ll

 ; RUN: opt < %s -basicaa -licm -S | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S %s | FileCheck %s
 
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-unknown-linux-gnu"
 
 ; Make sure we don't hoist the store out of the loop; %a would
 ; have the wrong value if f() unwinds
 
 define void @test1(i32* nocapture noalias %a, i1 zeroext %y) uwtable {

test/Transforms/LICM/scalar_promote.ll

 ; RUN: opt < %s -basicaa -tbaa -licm -S | FileCheck %s
+; RUN: opt -aa-pipeline=type-based-aa,basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S %s | FileCheck %s
 target datalayout = "E-p:64:64:64-a0:0:8-f32:32:32-f64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-v64:64:64-v128:128:128"
 
 @X = global i32 7   ; <i32*> [#uses=4]
 
 define void @test1(i32 %i) {
 Entry:
   br label %Loop
 ; CHECK-LABEL: @test1(

test/Transforms/LICM/speculate.ll

 ; RUN: opt -S -licm < %s | FileCheck %s
+; RUN: opt -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S %s | FileCheck %s
 
 ; UDiv is safe to speculate if the denominator is known non-zero.
 
 ; CHECK-LABEL: @safe_udiv(
 ; CHECK:      %div = udiv i64 %x, 2
 ; CHECK-NEXT: br label %for.body
 
 define void @safe_udiv(i64 %x, i64 %m, i64 %n, i32* %p, i64* %q) nounwind {

test/Transforms/LICM/volatile-alias.ll

 ; RUN: opt -basicaa -sroa -loop-rotate -licm -S < %s | FileCheck %s
+; RUN: opt -basicaa -sroa -loop-rotate %s | opt -aa-pipeline=basic-aa -passes='require<aa>,require<targetir>,require<scalar-evolution>,loop(licm)' -S | FileCheck %s
 ; The objects *p and *q are aliased to each other, but even though *q is
 ; volatile, *p can be considered invariant in the loop. Check if it is moved
 ; out of the loop.
 ; CHECK: load i32, i32* %p
 ; CHECK: for.body:
 ; CHECK: load volatile i32, i32* %q
 
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"