[PM][InstCombine] fixing omission of AliasAnalysis in new-pass-manager's version of InstCombine

[fedor.sergeev]

Summary

Passing AliasAnalysis results instead of nullptr appears to work just fine.
A couple new-pass-manager tests updated to align with new order of analyses.

[fedor.sergeev]

removing invalid comment

[chandlerc]

LGTM. It'd be great to add an instcombine test that uses new PM and requires AA to optimize as well. Can be in a follow-up commit if needed.

[skatkov]

lib/Transforms/InstCombine/InstructionCombining.cpp

3289 ↗

(On Diff #126840)

Chandler, may be you can answer this question: why do we need to preserve BasicAA if we already preserved AAManager?

[chandlerc]

lib/Transforms/InstCombine/InstructionCombining.cpp

3289 ↗

(On Diff #126840)

The AA manager just brokers access to the various different alias analyses. You still need to preserve the analyses themselves.

Now that the invalidation support is fully wired up, we could probably remove the need to explicitly preserve the AAManager and just have it remain preserved automatically if all of the AAs are preserved. But it isn't a big deal. The only cost to invalidating the AAManager is that we have to go and look up the N different alias analyses for the next pass.

[skatkov]

Thank you for explanation!

What is the valid way to preserve all alias analysis?

[fedor.sergeev]

adding a test

[fedor.sergeev]

Ahem, could somebody ok the test, pls?

[skatkov]

LGTM.

llvm/trunk/lib/Transforms/InstCombine/InstructionCombining.cpp

PreservedAnalyses InstCombinePass::run(Function &F,
                                        FunctionAnalysisManager &AM) {
   auto &AC = AM.getResult<AssumptionAnalysis>(F);
   auto &DT = AM.getResult<DominatorTreeAnalysis>(F);
   auto &TLI = AM.getResult<TargetLibraryAnalysis>(F);
   auto &ORE = AM.getResult<OptimizationRemarkEmitterAnalysis>(F);
 
   auto *LI = AM.getCachedResult<LoopAnalysis>(F);
 
-  // FIXME: The AliasAnalysis is not yet supported in the new pass manager
-  if (!combineInstructionsOverFunction(F, Worklist, nullptr, AC, TLI, DT, ORE,
+  auto *AA = &AM.getResult<AAManager>(F);
+  if (!combineInstructionsOverFunction(F, Worklist, AA, AC, TLI, DT, ORE,
                                        ExpensiveCombines, LI))
     // No changes, all analyses are preserved.
     return PreservedAnalyses::all();
 
   // Mark all the analyses that instcombine updates as preserved.
   PreservedAnalyses PA;
   PA.preserveSet<CFGAnalyses>();
   PA.preserve<AAManager>();
+  PA.preserve<BasicAA>();
   PA.preserve<GlobalsAA>();
   return PA;
 }
 
 void InstructionCombiningPass::getAnalysisUsage(AnalysisUsage &AU) const {
   AU.setPreservesCFG();
   AU.addRequired<AAResultsWrapperPass>();
   AU.addRequired<AssumptionCacheTracker>();

llvm/trunk/test/Other/new-pm-defaults.ll

 ; CHECK-O-NEXT: Running pass: CalledValuePropagationPass
 ; CHECK-O-NEXT: Running pass: GlobalOptPass
 ; CHECK-O-NEXT: Running pass: ModuleToFunctionPassAdaptor<{{.*}}PromotePass>
 ; CHECK-O-NEXT: Running pass: DeadArgumentEliminationPass
 ; CHECK-O-NEXT: Running pass: ModuleToFunctionPassAdaptor<{{.*}}PassManager{{.*}}>
 ; CHECK-O-NEXT: Starting llvm::Function pass manager run.
 ; CHECK-O-NEXT: Running pass: InstCombinePass
 ; CHECK-O-NEXT: Running analysis: OptimizationRemarkEmitterAnalysis
+; CHECK-O-NEXT: Running analysis: AAManager
 ; CHECK-EP-PEEPHOLE-NEXT: Running pass: NoOpFunctionPass
 ; CHECK-O-NEXT: Running pass: SimplifyCFGPass
 ; CHECK-O-NEXT: Finished llvm::Function pass manager run.
 ; CHECK-O-NEXT: Running pass: RequireAnalysisPass<{{.*}}GlobalsAA
 ; CHECK-O-NEXT: Running analysis: GlobalsAA
 ; CHECK-O-NEXT: Running analysis: CallGraphAnalysis
 ; CHECK-O-NEXT: Running pass: RequireAnalysisPass<{{.*}}ProfileSummaryAnalysis
 ; CHECK-O-NEXT: Running analysis: ProfileSummaryAnalysis
 ; CHECK-O-NEXT: Running pass: ModuleToPostOrderCGSCCPassAdaptor<{{.*}}LazyCallGraph{{.*}}>
 ; CHECK-O-NEXT: Running analysis: InnerAnalysisManagerProxy
 ; CHECK-O-NEXT: Running analysis: LazyCallGraphAnalysis
 ; CHECK-O-NEXT: Starting CGSCC pass manager run.
 ; CHECK-O-NEXT: Running pass: InlinerPass
 ; CHECK-O-NEXT: Running analysis: OuterAnalysisManagerProxy<{{.*}}LazyCallGraph{{.*}}>
 ; CHECK-O-NEXT: Running pass: PostOrderFunctionAttrsPass
 ; CHECK-O-NEXT: Running analysis: FunctionAnalysisManagerCGSCCProxy
-; CHECK-O-NEXT: Running analysis: AAManager
 ; CHECK-O3-NEXT: Running pass: ArgumentPromotionPass
 ; CHECK-O-NEXT: Running pass: CGSCCToFunctionPassAdaptor<{{.*}}PassManager{{.*}}>
 ; CHECK-O-NEXT: Starting llvm::Function pass manager run.
 ; CHECK-O-NEXT: Running pass: SROA
 ; CHECK-O-NEXT: Running pass: EarlyCSEPass
 ; CHECK-O-NEXT: Running analysis: MemorySSAAnalysis
 ; CHECK-O-NEXT: Running pass: SpeculativeExecutionPass
 ; CHECK-O-NEXT: Running pass: JumpThreadingPass

llvm/trunk/test/Other/new-pm-thinlto-defaults.ll

 ; CHECK-O-NEXT: Running pass: CalledValuePropagationPass
 ; CHECK-O-NEXT: Running pass: GlobalOptPass
 ; CHECK-O-NEXT: Running pass: ModuleToFunctionPassAdaptor<{{.*}}PromotePass>
 ; CHECK-O-NEXT: Running pass: DeadArgumentEliminationPass
 ; CHECK-O-NEXT: Running pass: ModuleToFunctionPassAdaptor<{{.*}}PassManager{{.*}}>
 ; CHECK-O-NEXT: Starting llvm::Function pass manager run.
 ; CHECK-O-NEXT: Running pass: InstCombinePass
 ; CHECK-PRELINK-O-NEXT: Running analysis: OptimizationRemarkEmitterAnalysis
+; CHECK-O-NEXT: Running analysis: AAManager
 ; CHECK-O-NEXT: Running pass: SimplifyCFGPass
 ; CHECK-O-NEXT: Finished llvm::Function pass manager run.
 ; CHECK-O-NEXT: Running pass: RequireAnalysisPass<{{.*}}GlobalsAA
 ; CHECK-O-NEXT: Running analysis: GlobalsAA
 ; CHECK-O-NEXT: Running analysis: CallGraphAnalysis
 ; CHECK-O-NEXT: Running pass: RequireAnalysisPass<{{.*}}ProfileSummaryAnalysis
 ; CHECK-PRELINK-O-NEXT: Running analysis: ProfileSummaryAnalysis
 ; CHECK-O-NEXT: Running pass: ModuleToPostOrderCGSCCPassAdaptor<{{.*}}LazyCallGraph{{.*}}>
 ; CHECK-O-NEXT: Running analysis: InnerAnalysisManagerProxy
 ; CHECK-O-NEXT: Running analysis: LazyCallGraphAnalysis
 ; CHECK-O-NEXT: Starting CGSCC pass manager run.
 ; CHECK-O-NEXT: Running pass: InlinerPass
 ; CHECK-O-NEXT: Running analysis: OuterAnalysisManagerProxy<{{.*}}LazyCallGraph{{.*}}>
 ; CHECK-O-NEXT: Running pass: PostOrderFunctionAttrsPass
 ; CHECK-O-NEXT: Running analysis: FunctionAnalysisManagerCGSCCProxy
-; CHECK-O-NEXT: Running analysis: AAManager
 ; CHECK-O3-NEXT: Running pass: ArgumentPromotionPass
 ; CHECK-O-NEXT: Running pass: CGSCCToFunctionPassAdaptor<{{.*}}PassManager{{.*}}>
 ; CHECK-O-NEXT: Starting llvm::Function pass manager run.
 ; CHECK-O-NEXT: Running pass: SROA
 ; CHECK-O-NEXT: Running pass: EarlyCSEPass
 ; CHECK-O-NEXT: Running analysis: MemorySSAAnalysis
 ; CHECK-O-NEXT: Running pass: SpeculativeExecutionPass
 ; CHECK-O-NEXT: Running pass: JumpThreadingPass

llvm/trunk/test/Transforms/InstCombine/store-load-unaliased-gep.ll

+; RUN: opt -instcombine %s -S 2>&1 | FileCheck %s
+; RUN: opt -aa-pipeline=basic-aa -passes=instcombine %s -S 2>&1 | FileCheck %s
+
+; Checking successful store-load optimization of array length.
+; Function below should deduce just to "return length".
+; Doable only if instcombine has access to alias-analysis.
+
+define i32 @test1(i32 %length) {
+; CHECK-LABEL: entry:
+entry:
+  %array = alloca i32, i32 2
+  ; CHECK-NOT: %array
+
+  %length_gep = getelementptr inbounds i32, i32 * %array, i32 0
+  %value_gep = getelementptr inbounds i32, i32 * %array, i32 1
+  store i32 %length, i32 * %length_gep
+  store i32 0, i32 * %value_gep
+  %loaded_length = load i32, i32 * %length_gep
+  ; CHECK-NOT: %loaded_length = load i32
+
+  ret i32 %loaded_length
+  ; CHECK: ret i32 %length
+}