[Pipelines] Add LoopSink and DivRemPairs to LTO post-link pipeline

[nikic]

Summary

As pointed out in D148010, these passes are missing from the LTO post-link pipeline. They are present in the pre-link pipeline, but LoopSink is completely useless there (it will always be fully undone by LICM post-link) and DivRemPairs is mostly useless (I believe most of what it does will be undone by InstCombine).

I've not added RelLookupTableConverterPass, because it's also disabled in the LTO pre-link pipeline, with a comment that there is an unresolved issue with full LTO.

Compile-time impact of the extra passes is minimal: http://llvm-compile-time-tracker.com/compare.php?from=09ba7b605327812cfcec4f3c01d7fc232dee651d&to=52d546b4fe89ad66011ca9e3a8f2f6dfe4b4ac2f&stat=instructions:u Of course, LoopSink will have a larger impact in PGO builds.

[aeubanks]

llvm/lib/Passes/PassBuilderPipelines.cpp

1836

this runs after the function passes in the module optimization pipeline

llvm/lib/Passes/PassBuilderPipelines.cpp

PassBuilder::buildLTODefaultPipeline(OptimizationLevel Level,
   // CFI is disabled.
   MPM.addPass(LowerTypeTestsPass(ExportSummary, nullptr));
   // Run a second time to clean up any type tests left behind by WPD for use
   // in ICP (which is performed earlier than this in the regular LTO pipeline).
   MPM.addPass(LowerTypeTestsPass(nullptr, nullptr, true));
 
   // Enable splitting late in the FullLTO post-link pipeline.
   if (EnableHotColdSplit)
     MPM.addPass(HotColdSplittingPass());
 
   // Add late LTO optimization passes.
+  FunctionPassManager LateFPM;
+
+  // LoopSink pass sinks instructions hoisted by LICM, which serves as a
+  // canonicalization pass that enables other optimizations. As a result,
+  // LoopSink pass needs to be a very late IR pass to avoid undoing LICM
+  // result too early.
+  LateFPM.addPass(LoopSinkPass());
+
+  // This hoists/decomposes div/rem ops. It should run after other sink/hoist
+  // passes to avoid re-sinking, but before SimplifyCFG because it can allow
+  // flattening of blocks.
+  LateFPM.addPass(DivRemPairsPass());
+
   // Delete basic blocks, which optimization passes may have killed.
-  MPM.addPass(createModuleToFunctionPassAdaptor(SimplifyCFGPass(
+  LateFPM.addPass(SimplifyCFGPass(
       SimplifyCFGOptions().convertSwitchRangeToICmp(true).hoistCommonInsts(
-          true))));
+          true)));
+  MPM.addPass(createModuleToFunctionPassAdaptor(std::move(LateFPM)));
 
   // Drop bodies of available eternally objects to improve GlobalDCE.
   MPM.addPass(EliminateAvailableExternallyPass());
 
   // Now that we have optimized the program, discard unreachable functions.
   MPM.addPass(GlobalDCEPass());
 
   if (PTO.MergeFunctions)

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

 ; CHECK-OS-NEXT: Running pass: SLPVectorizerPass on foo
 ; CHECK-O23SZ-NEXT: Running pass: VectorCombinePass on foo
 ; CHECK-O23SZ-NEXT: Running pass: AlignmentFromAssumptionsPass on foo
 ; CHECK-O23SZ-NEXT: Running pass: InstCombinePass on foo
 ; CHECK-EP-Peephole-NEXT: Running pass: NoOpFunctionPass on foo
 ; CHECK-O23SZ-NEXT: Running pass: JumpThreadingPass on foo
 ; CHECK-O23SZ-NEXT: Running pass: LowerTypeTestsPass
 ; CHECK-O-NEXT: Running pass: LowerTypeTestsPass
+; CHECK-O23SZ-NEXT: Running pass: LoopSink
+; CHECK-O23SZ-NEXT: Running pass: DivRemPairs
 ; CHECK-O23SZ-NEXT: Running pass: SimplifyCFGPass
 ; CHECK-O23SZ-NEXT: Running pass: EliminateAvailableExternallyPass
 ; CHECK-O23SZ-NEXT: Running pass: GlobalDCEPass
 ; CHECK-O23SZ-NEXT: Running pass: CGProfilePass
 ; CHECK-EP-NEXT: Running pass: NoOpModulePass
 ; CHECK-O-NEXT: Running pass: AnnotationRemarksPass on foo
 ; CHECK-O-NEXT: Running pass: PrintModulePass