[Passes] Run GlobalsAA before LICM during LTO in new PM.

[fhahn]

Summary

This patch adjusts the LTO pipeline in the new PM to run GlobalsAA
before LICM to match the legacy PM.

This fixes a regression where the new PM failed to vectorize loops that
require hoisting/sinking by LICM depending on GlobalsAA info.

[aeubanks]

lgtm

[nikic]

These changes don't look related to me, can you please land them separately?

[fhahn]

In D102345#2754939, @nikic wrote:

These changes don't look related to me, can you please land them separately?

Yeah, they are only loosely related through the test case, which requires both. I'll submit the MemorySSA change separately first.

[fhahn]

Strip away memoryssa change.

[vvereschaka]

Hi @fhahn,

these changes breaks the new-pm-lto-defaults.ll test on the cross win-x-arm linux cross builders:

• https://lab.llvm.org/buildbot/#/builders/119/builds/3775/steps/8/logs/FAIL__LLVM__new-pm-lto-defaults_ll
• https://lab.llvm.org/buildbot/#/builders/60/builds/3287/steps/8/logs/FAIL__LLVM__new-pm-lto-defaults_ll

The problem is the test expects the following pattern: Running pass: InvalidateAnalysisPass<llvm::AAManager> on foo
but it gets Running pass: InvalidateAnalysisPass<class llvm::AAManager> on foo from the input.

Would you take care of it?

[aeubanks]

Should be fixed with ab6a609d961181bbac0d1ea0a830061049990020.

[vvereschaka]

Thank you @aeubanks, got it.

llvm/lib/Passes/PassBuilder.cpp

PassBuilder::buildLTODefaultPipeline(OptimizationLevel Level,
   // link-time inlining, and visibility of nocapture attribute.
   FPM.addPass(TailCallElimPass());
 
   // Run a few AA driver optimizations here and now to cleanup the code.
   MPM.addPass(createModuleToFunctionPassAdaptor(std::move(FPM)));
 
   MPM.addPass(
       createModuleToPostOrderCGSCCPassAdaptor(PostOrderFunctionAttrsPass()));
-  // FIXME: here we run IP alias analysis in the legacy PM.
 
-  FunctionPassManager MainFPM;
+  // Require the GlobalsAA analysis for the module so we can query it within
+  // MainFPM.
+  MPM.addPass(RequireAnalysisPass<GlobalsAA, Module>());
+  // Invalidate AAManager so it can be recreated and pick up the newly available
+  // GlobalsAA.
+  MPM.addPass(
+      createModuleToFunctionPassAdaptor(InvalidateAnalysisPass<AAManager>()));
 
+  FunctionPassManager MainFPM;
   MainFPM.addPass(createFunctionToLoopPassAdaptor(
-      LICMPass(PTO.LicmMssaOptCap, PTO.LicmMssaNoAccForPromotionCap)));
+      LICMPass(PTO.LicmMssaOptCap, PTO.LicmMssaNoAccForPromotionCap),
+      EnableMSSALoopDependency, /*UseBlockFrequencyInfo=*/true));
 
   if (RunNewGVN)
     MainFPM.addPass(NewGVNPass());
   else
     MainFPM.addPass(GVN());
 
   // Remove dead memcpy()'s.
   MainFPM.addPass(MemCpyOptPass());

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

 ; CHECK-O23SZ-NEXT: Running pass: GlobalDCEPass
 ; CHECK-O23SZ-NEXT: Running pass: InstCombinePass
 ; CHECK-EP-Peephole-NEXT: Running pass: NoOpFunctionPass
 ; CHECK-O23SZ-NEXT: Running pass: JumpThreadingPass
 ; CHECK-O23SZ-NEXT: Running analysis: LazyValueAnalysis
 ; CHECK-O23SZ-NEXT: Running pass: SROA on foo
 ; CHECK-O23SZ-NEXT: Running pass: TailCallElimPass on foo
 ; CHECK-O23SZ-NEXT: Running pass: PostOrderFunctionAttrsPass on (foo)
+; CHECK-O23SZ-NEXT: Running pass: RequireAnalysisPass<llvm::GlobalsAA, llvm::Module> on [module]
+; CHECK-O23SZ-NEXT: Running analysis: GlobalsAA on [module]
+; CHECK-O23SZ-NEXT: Running pass: InvalidateAnalysisPass<llvm::AAManager> on foo
+; CHECK-O23SZ-NEXT: Invalidating analysis: AAManager on foo
 ; CHECK-O23SZ-NEXT: Running pass: LoopSimplifyPass on foo
 ; CHECK-O23SZ-NEXT: Running analysis: LoopAnalysis on foo
 ; CHECK-O23SZ-NEXT: Running pass: LCSSAPass on foo
+; CHECK-O23SZ-NEXT: Running analysis: MemorySSAAnalysis on foo
+; CHECK-O23SZ-NEXT: Running analysis: AAManager on foo
 ; CHECK-O23SZ-NEXT: Running analysis: ScalarEvolutionAnalysis on foo
 ; CHECK-O23SZ-NEXT: Running analysis: InnerAnalysisManagerProxy
 ; CHECK-O23SZ-NEXT: Running pass: LICMPass on Loop
 ; CHECK-O23SZ-NEXT: Running pass: GVN on foo
 ; CHECK-O23SZ-NEXT: Running analysis: MemoryDependenceAnalysis on foo
 ; CHECK-O23SZ-NEXT: Running analysis: PhiValuesAnalysis on foo
 ; CHECK-O23SZ-NEXT: Running pass: MemCpyOptPass on foo
-; CHECK-O23SZ-NEXT: Running analysis: MemorySSAAnalysis on foo
 ; CHECK-O23SZ-NEXT: Running pass: DSEPass on foo
 ; CHECK-O23SZ-NEXT: Running analysis: PostDominatorTreeAnalysis on foo
 ; CHECK-O23SZ-NEXT: Running pass: MergedLoadStoreMotionPass on foo
 ; CHECK-O23SZ-NEXT: Running pass: LoopSimplifyPass on foo
 ; CHECK-O23SZ-NEXT: Running pass: LCSSAPass on foo
 ; CHECK-O23SZ-NEXT: Running pass: IndVarSimplifyPass on Loop
 ; CHECK-O23SZ-NEXT: Running pass: LoopDeletionPass on Loop
 ; CHECK-O23SZ-NEXT: Running pass: LoopFullUnrollPass on Loop

llvm/test/Transforms/PhaseOrdering/AArch64/globals-aa-required-for-vectorization.ll

 ; RUN: opt -passes='lto<O3>' -S %s | FileCheck %s
-; RUN: opt -std-link-opts -enable-new-pm=false -S %s | FileCheck --check-prefix=LEGACY %s
+; RUN: opt -std-link-opts -enable-new-pm=false -S %s | FileCheck %s
 
 target triple = "arm64e-apple-darwin"
 
 @A = external unnamed_addr global i8**, align 8
 @B = external unnamed_addr global i8**, align 8
 @C = internal unnamed_addr global i32 0, align 4
 @D = external unnamed_addr global i32, align 4
 
 ; CHECK-LABEL: @fn
-; CHECK-NOT: vector.body:
-;
-; LEGACY-LABEL: @fn
-; LEGACY: vector.body:
+; CHECK: vector.body:
 ;
 define void @fn() {
 entry:
   %v.D = load i32, i32* @D, align 4
   store i32 %v.D, i32* @C, align 4
   call void @clobber()
 
   %v.B = load i8**, i8*** @B, align 8