[newpm] Add an OptimizerLast EP

[philip.pfaffe]

Summary

It turns out that we need an OptimizerLast PassBuilder extension point
after all. I missed the relevance of this EP the first time. By legacy PM magic,
function passes added at this EP get added to the last _Function_ PM, which is a
feature we lost when dropping this EP for the new PM.

[fedor.sergeev]

LGTM (though I have no idea what legacy does for this :) )

tools/opt/NewPMDriver.cpp

95

Unrelated to your fix, shouldnt this read "of the MODULE pass pipeline" ?

[philip.pfaffe]

A key difference between this and legacy is at O0, this EP won't be triggered.

[fedor.sergeev]

In D54374#1294134, @philip.pfaffe wrote:

A key difference between this and legacy is at O0, this EP won't be triggered.

Makes sense to mention this in commit message...

[chandlerc]

Generally feel free to submit w/ comment and commit message updates.

(But also I want to explicitly say you don't need to wait for me on this kind of patch, very happy w/ Fedor's review.)

include/llvm/Passes/PassBuilder.h

501–502

Also mention the O0 thing here?

include/llvm/Passes/PassBuilder.h

public:
   /// This extension point allows adding optimization once at the start of the
   /// pipeline. This does not apply to 'backend' compiles (LTO and ThinLTO
   /// link-time pipelines).
   void registerPipelineStartEPCallback(
       const std::function<void(ModulePassManager &)> &C) {
     PipelineStartEPCallbacks.push_back(C);
   }
 
+  /// Register a callback for a default optimizer pipeline extension point
+  ///
+  /// This extension point allows adding optimizations at the very end of the
+  /// function optimization pipeline.
+  void registerOptimizerLastEPCallback(
+      const std::function<void(FunctionPassManager &, OptimizationLevel)> &C) {
+    OptimizerLastEPCallbacks.push_back(C);
+  }
+
   /// Register a callback for parsing an AliasAnalysis Name to populate
   /// the given AAManager \p AA
   void registerParseAACallback(
       const std::function<bool(StringRef Name, AAManager &AA)> &C) {
     AAParsingCallbacks.push_back(C);
   }
 
   /// {{@ Register callbacks for analysis registration with this PassBuilder
private:
   SmallVector<std::function<void(LoopPassManager &, OptimizationLevel)>, 2>
       LoopOptimizerEndEPCallbacks;
   SmallVector<std::function<void(FunctionPassManager &, OptimizationLevel)>, 2>
       ScalarOptimizerLateEPCallbacks;
   SmallVector<std::function<void(CGSCCPassManager &, OptimizationLevel)>, 2>
       CGSCCOptimizerLateEPCallbacks;
   SmallVector<std::function<void(FunctionPassManager &, OptimizationLevel)>, 2>
       VectorizerStartEPCallbacks;
+  SmallVector<std::function<void(FunctionPassManager &, OptimizationLevel)>, 2>
+      OptimizerLastEPCallbacks;
   // Module callbacks
   SmallVector<std::function<void(ModulePassManager &)>, 2>
       PipelineStartEPCallbacks;
   SmallVector<std::function<void(ModuleAnalysisManager &)>, 2>
       ModuleAnalysisRegistrationCallbacks;
   SmallVector<std::function<bool(StringRef, ModulePassManager &,
                                  ArrayRef<PipelineElement>)>,
               2>

lib/Passes/PassBuilder.cpp

PassBuilder::buildModuleOptimizationPipeline(OptimizationLevel Level,
   // resulted in single-entry-single-exit or empty blocks. Clean up the CFG.
   OptimizePM.addPass(SimplifyCFGPass());
 
   // Optimize PHIs by speculating around them when profitable. Note that this
   // pass needs to be run after any PRE or similar pass as it is essentially
   // inserting redudnancies into the progrem. This even includes SimplifyCFG.
   OptimizePM.addPass(SpeculateAroundPHIsPass());
 
+  for (auto &C : OptimizerLastEPCallbacks)
+    C(OptimizePM, Level);
+
   // Add the core optimizing pipeline.
   MPM.addPass(createModuleToFunctionPassAdaptor(std::move(OptimizePM)));
 
   MPM.addPass(CGProfilePass());
 
   // Now we need to do some global optimization transforms.
   // FIXME: It would seem like these should come first in the optimization
   // pipeline and maybe be the bottom of the canonicalization pipeline? Weird

test/Other/new-pm-defaults.ll

 ; RUN:     -passes='default<O3>' -S  %s 2>&1 \
 ; RUN:     | FileCheck %s --check-prefix=CHECK-O --check-prefix=CHECK-O3 \
 ; RUN:     --check-prefix=CHECK-EP-PIPELINE-START
 ; RUN: opt -disable-verify -debug-pass-manager \
 ; RUN:     -passes-ep-pipeline-start='no-op-module' \
 ; RUN:     -passes='lto-pre-link<O3>' -S  %s 2>&1 \
 ; RUN:     | FileCheck %s --check-prefix=CHECK-O --check-prefix=CHECK-O3 \
 ; RUN:     --check-prefix=CHECK-EP-PIPELINE-START
+; RUN: opt -disable-verify -debug-pass-manager \
+; RUN:     -passes-ep-optimizer-last='no-op-function' \
+; RUN:     -passes='default<O3>' -S  %s 2>&1 \
+; RUN:     | FileCheck %s --check-prefix=CHECK-O --check-prefix=CHECK-O3 \
+; RUN:     --check-prefix=CHECK-EP-OPTIMIZER-LAST
 
 ; CHECK-O: Running analysis: PassInstrumentationAnalysis
 ; CHECK-O-NEXT: Starting llvm::Module pass manager run.
 ; CHECK-O-NEXT: Running pass: PassManager<{{.*}}Module{{.*}}>
 ; CHECK-O-NEXT: Starting llvm::Module pass manager run.
 ; CHECK-O-NEXT: Running pass: ForceFunctionAttrsPass
 ; CHECK-EP-PIPELINE-START-NEXT: Running pass: NoOpModulePass
 ; CHECK-O-NEXT: Running pass: PassManager<{{.*}}Module{{.*}}>

 ; CHECK-O-NEXT: Running pass: LCSSAPass
 ; CHECK-O-NEXT: Finished llvm::Function pass manager run.
 ; CHECK-O-NEXT: Running pass: AlignmentFromAssumptionsPass
 ; CHECK-O-NEXT: Running pass: LoopSinkPass
 ; CHECK-O-NEXT: Running pass: InstSimplifyPass
 ; CHECK-O-NEXT: Running pass: DivRemPairsPass
 ; CHECK-O-NEXT: Running pass: SimplifyCFGPass
 ; CHECK-O-NEXT: Running pass: SpeculateAroundPHIsPass
+; CHECK-EP-OPTIMIZER-LAST: Running pass: NoOpFunctionPass
 ; CHECK-O-NEXT: Finished llvm::Function pass manager run.
 ; CHECK-O-NEXT: Running pass: CGProfilePass
 ; CHECK-O-NEXT: Running pass: GlobalDCEPass
 ; CHECK-O-NEXT: Running pass: ConstantMergePass
 ; CHECK-O-NEXT: Finished llvm::Module pass manager run.
 ; CHECK-O-NEXT: Finished llvm::Module pass manager run.
 ; CHECK-O-NEXT: Running pass: PrintModulePass
 ;

tools/opt/NewPMDriver.cpp

static cl::opt<std::string> CGSCCOptimizerLateEPPipeline(
     cl::Hidden);
 static cl::opt<std::string> VectorizerStartEPPipeline(
     "passes-ep-vectorizer-start",
     cl::desc("A textual description of the function pass pipeline inserted at "
              "the VectorizerStart extension point into default pipelines"),
     cl::Hidden);
 static cl::opt<std::string> PipelineStartEPPipeline(
     "passes-ep-pipeline-start",
     cl::desc("A textual description of the function pass pipeline inserted at "
              "the PipelineStart extension point into default pipelines"),
     cl::Hidden);
+static cl::opt<std::string> OptimizerLastEPPipeline(
+    "passes-ep-optimizer-last",
+    cl::desc("A textual description of the function pass pipeline inserted at "
+             "the OptimizerLast extension point into default pipelines"),
+    cl::Hidden);
+
 enum PGOKind { NoPGO, InstrGen, InstrUse, SampleUse };
 static cl::opt<PGOKind> PGOKindFlag(
     "pgo-kind", cl::init(NoPGO), cl::Hidden,
     cl::desc("The kind of profile guided optimization"),
     cl::values(clEnumValN(NoPGO, "nopgo", "Do not use PGO."),
                clEnumValN(InstrGen, "new-pm-pgo-instr-gen-pipeline",
                           "Instrument the IR to generate profile."),
                clEnumValN(InstrUse, "new-pm-pgo-instr-use-pipeline",
PB.registerVectorizerStartEPCallback([&PB, VerifyEachPass, DebugLogging](
                            DebugLogging);
     });
   if (tryParsePipelineText<ModulePassManager>(PB, PipelineStartEPPipeline))
     PB.registerPipelineStartEPCallback(
         [&PB, VerifyEachPass, DebugLogging](ModulePassManager &PM) {
           PB.parsePassPipeline(PM, PipelineStartEPPipeline, VerifyEachPass,
                                DebugLogging);
         });
+  if (tryParsePipelineText<FunctionPassManager>(PB, OptimizerLastEPPipeline))
+    PB.registerOptimizerLastEPCallback(
+        [&PB, VerifyEachPass, DebugLogging](FunctionPassManager &PM,
+                                            PassBuilder::OptimizationLevel) {
+          PB.parsePassPipeline(PM, OptimizerLastEPPipeline, VerifyEachPass,
+                               DebugLogging);
+        });
 }
 
 #ifdef LINK_POLLY_INTO_TOOLS
 namespace polly {
 void RegisterPollyPasses(PassBuilder &);
 }
 #endif