[NFC][PowerPC] Custom PowerPC specific machine-scheduler

[steven.zhang]

Summary

This patch lays the groundwork for extending the generic machine scheduler by providing a PPC-specific implementation. There are no functional changes as this is an incremental patch that simply provides the necessary overrides which just encapsulate the behavior of the generic scheduler. Subsequent patches will add specific behavior.

[steven.zhang]

gentle ping ..

[jsji]

LGTM, except for some minor comments change.
Thanks for extending this, it will allow us to tune MachineScheduler specifically for PowerPC going forward.

llvm/lib/Target/PowerPC/PPCMachineScheduler.h

1 ↗

(On Diff #190372)

Why an empty first line?

21 ↗

(On Diff #190372)

Maybe add "pre RA" in comment too, as "post RA" in below?

21 ↗

(On Diff #190372)

Doxgen comment instead? //' -> ///'

28 ↗

(On Diff #190372)

Doxgen comment instead?

[steven.zhang]

Thank you for the review comments. I have addressed all the comments.

llvm/trunk/lib/Target/PowerPC/CMakeLists.txt

add_llvm_target(PowerPCCodeGen
   PPCISelDAGToDAG.cpp
   PPCISelLowering.cpp
   PPCEarlyReturn.cpp
   PPCFastISel.cpp
   PPCFrameLowering.cpp
   PPCLoopPreIncPrep.cpp
   PPCMCInstLower.cpp
   PPCMachineFunctionInfo.cpp
+  PPCMachineScheduler.cpp
   PPCMIPeephole.cpp
   PPCRegisterInfo.cpp
   PPCQPXLoadSplat.cpp
   PPCSubtarget.cpp
   PPCTargetMachine.cpp
   PPCTargetObjectFile.cpp
   PPCTargetTransformInfo.cpp
   PPCTOCRegDeps.cpp

llvm/trunk/lib/Target/PowerPC/PPC.td

 def FeatureLongCall : SubtargetFeature<"longcall", "UseLongCalls", "true",
                                        "Always use indirect calls">;
 def FeatureHTM : SubtargetFeature<"htm", "HasHTM", "true",
                                   "Enable Hardware Transactional Memory instructions">;
 def FeatureMFTB   : SubtargetFeature<"", "FeatureMFTB", "true",
                                         "Implement mftb using the mfspr instruction">;
 def FeatureFusion : SubtargetFeature<"fusion", "HasFusion", "true",
                                      "Target supports add/load integer fusion.">;
+def FeaturePPCPreRASched:
+  SubtargetFeature<"ppc-prera-sched", "UsePPCPreRASchedStrategy", "true",
+                   "Use PowerPC pre-RA scheduling strategy">;
+def FeaturePPCPostRASched:
+  SubtargetFeature<"ppc-postra-sched", "UsePPCPostRASchedStrategy", "true",
+                   "Use PowerPC post-RA scheduling strategy">;
 def FeatureFloat128 :
   SubtargetFeature<"float128", "HasFloat128", "true",
                    "Enable the __float128 data type for IEEE-754R Binary128.",
                    [FeatureVSX]>;
 def FeaturePOPCNTD   : SubtargetFeature<"popcntd","HasPOPCNTD",
                                         "POPCNTD_Fast",
                                         "Enable the popcnt[dw] instructions">;
 // Note that for the a2/a2q processor models we should not use popcnt[dw] by
def ProcessorFeatures {
   list<SubtargetFeature> Power8SpecificFeatures =
       [DirectivePwr8, FeatureP8Altivec, FeatureP8Vector, FeatureP8Crypto,
        FeatureHTM, FeatureDirectMove, FeatureICBT, FeaturePartwordAtomic,
        FeatureFusion];
   list<SubtargetFeature> Power8FeatureList =
       !listconcat(Power7FeatureList, Power8SpecificFeatures);
   list<SubtargetFeature> Power9SpecificFeatures =
       [DirectivePwr9, FeatureP9Altivec, FeatureP9Vector, FeatureISA3_0,
-       FeatureVectorsUseTwoUnits];
+       FeatureVectorsUseTwoUnits, FeaturePPCPreRASched, FeaturePPCPostRASched];
   list<SubtargetFeature> Power9FeatureList =
       !listconcat(Power8FeatureList, Power9SpecificFeatures);
 }
 
 // Note: Future features to add when support is extended to more
 // recent ISA levels:
 //
 // DFP          p6, p6x, p7        decimal floating-point instructions

llvm/trunk/lib/Target/PowerPC/PPCMachineScheduler.h

+//===- PPCMachineScheduler.h - Custom PowerPC MI scheduler --*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// Custom PowerPC MI scheduler.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIB_TARGET_POWERPC_POWERPCMACHINESCHEDULER_H
+#define LLVM_LIB_TARGET_POWERPC_POWERPCMACHINESCHEDULER_H
+
+#include "llvm/CodeGen/MachineScheduler.h"
+
+namespace llvm {
+
+/// A MachineSchedStrategy implementation for PowerPC pre RA scheduling.
+class PPCPreRASchedStrategy : public GenericScheduler {
+public:
+  PPCPreRASchedStrategy(const MachineSchedContext *C) :
+    GenericScheduler(C) {}
+};
+
+/// A MachineSchedStrategy implementation for PowerPC post RA scheduling.
+class PPCPostRASchedStrategy : public PostGenericScheduler {
+public:
+  PPCPostRASchedStrategy(const MachineSchedContext *C) :
+    PostGenericScheduler(C) {}
+
+protected:
+  void initialize(ScheduleDAGMI *Dag) override;
+  SUnit *pickNode(bool &IsTopNode) override;
+  void enterMBB(MachineBasicBlock *MBB) override;
+  void leaveMBB() override;
+};
+
+} // end namespace llvm
+
+#endif // LLVM_LIB_TARGET_POWERPC_POWERPCMACHINESCHEDULER_H

llvm/trunk/lib/Target/PowerPC/PPCMachineScheduler.cpp

+//===- PPCMachineScheduler.cpp - MI Scheduler for PowerPC -------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+#include "PPCMachineScheduler.h"
+using namespace llvm;
+
+void PPCPostRASchedStrategy::enterMBB(MachineBasicBlock *MBB) {
+  // Custom PPC PostRA specific behavior here.
+  PostGenericScheduler::enterMBB(MBB);
+}
+
+void PPCPostRASchedStrategy::leaveMBB() {
+  // Custom PPC PostRA specific behavior here.
+  PostGenericScheduler::leaveMBB();
+}
+
+void PPCPostRASchedStrategy::initialize(ScheduleDAGMI *Dag) {
+  // Custom PPC PostRA specific initialization here.
+  PostGenericScheduler::initialize(Dag);
+}
+
+SUnit *PPCPostRASchedStrategy::pickNode(bool &IsTopNode) {
+  // Custom PPC PostRA specific scheduling here.
+  return PostGenericScheduler::pickNode(IsTopNode);
+}
+

llvm/trunk/lib/Target/PowerPC/PPCSubtarget.h

protected:
   bool HasDirectMove;
   bool HasHTM;
   bool HasFusion;
   bool HasFloat128;
   bool IsISA3_0;
   bool UseLongCalls;
   bool SecurePlt;
   bool VectorsUseTwoUnits;
+  bool UsePPCPreRASchedStrategy;
+  bool UsePPCPostRASchedStrategy;
 
   POPCNTDKind HasPOPCNTD;
 
   /// When targeting QPX running a stock PPC64 Linux kernel where the stack
   /// alignment has not been changed, we need to keep the 16-byte alignment
   /// of the stack.
   bool IsQPXStackUnaligned;
 
public:
   bool vectorsUseTwoUnits() const {return VectorsUseTwoUnits; }
   bool isE500() const { return IsE500; }
   bool isFeatureMFTB() const { return FeatureMFTB; }
   bool isDeprecatedDST() const { return DeprecatedDST; }
   bool hasICBT() const { return HasICBT; }
   bool hasInvariantFunctionDescriptors() const {
     return HasInvariantFunctionDescriptors;
   }
+  bool usePPCPreRASchedStrategy() const { return UsePPCPreRASchedStrategy; }
+  bool usePPCPostRASchedStrategy() const { return UsePPCPostRASchedStrategy; }
   bool hasPartwordAtomics() const { return HasPartwordAtomics; }
   bool hasDirectMove() const { return HasDirectMove; }
 
   bool isQPXStackUnaligned() const { return IsQPXStackUnaligned; }
   unsigned getPlatformStackAlignment() const {
     if ((hasQPX() || isBGQ()) && !isQPXStackUnaligned())
       return 32;
 

llvm/trunk/lib/Target/PowerPC/PPCSubtarget.cpp

void PPCSubtarget::initializeEnvironment() {
   IsQPXStackUnaligned = false;
   HasHTM = false;
   HasFusion = false;
   HasFloat128 = false;
   IsISA3_0 = false;
   UseLongCalls = false;
   SecurePlt = false;
   VectorsUseTwoUnits = false;
+  UsePPCPreRASchedStrategy = false;
+  UsePPCPostRASchedStrategy = false;
 
   HasPOPCNTD = POPCNTD_Unavailable;
 }
 
 void PPCSubtarget::initSubtargetFeatures(StringRef CPU, StringRef FS) {
   // Determine default and user specified characteristics
   std::string CPUName = CPU;
   if (CPUName.empty() || CPU == "generic") {

llvm/trunk/lib/Target/PowerPC/PPCTargetMachine.cpp

 //===-- PPCTargetMachine.cpp - Define TargetMachine for PowerPC -----------===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 //
 // Top-level implementation for the PowerPC target.
 //
 //===----------------------------------------------------------------------===//
 
 #include "PPCTargetMachine.h"
 #include "MCTargetDesc/PPCMCTargetDesc.h"
 #include "PPC.h"
+#include "PPCMachineScheduler.h"
 #include "PPCSubtarget.h"
 #include "PPCTargetObjectFile.h"
 #include "PPCTargetTransformInfo.h"
 #include "llvm/ADT/Optional.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/StringRef.h"
 #include "llvm/ADT/Triple.h"
 #include "llvm/Analysis/TargetTransformInfo.h"
if (CM) {
     return *CM;
   }
   if (!TT.isOSDarwin() && !JIT &&
       (TT.getArch() == Triple::ppc64 || TT.getArch() == Triple::ppc64le))
     return CodeModel::Medium;
   return CodeModel::Small;
 }
 
+
+static ScheduleDAGInstrs *createPPCMachineScheduler(MachineSchedContext *C) {
+  ScheduleDAGMILive *DAG =
+    new ScheduleDAGMILive(C, llvm::make_unique<PPCPreRASchedStrategy>(C));
+  // add DAG Mutations here.
+  DAG->addMutation(createCopyConstrainDAGMutation(DAG->TII, DAG->TRI));
+  return DAG;
+}
+
+static ScheduleDAGInstrs *createPPCPostMachineScheduler(
+  MachineSchedContext *C) {
+  ScheduleDAGMI *DAG =
+    new ScheduleDAGMI(C, llvm::make_unique<PPCPostRASchedStrategy>(C), true);
+  // add DAG Mutations here.
+  return DAG;
+}
+
 // The FeatureString here is a little subtle. We are modifying the feature
 // string with what are (currently) non-function specific overrides as it goes
 // into the LLVMTargetMachine constructor and then using the stored value in the
 // Subtarget constructor below it.
 PPCTargetMachine::PPCTargetMachine(const Target &T, const Triple &TT,
                                    StringRef CPU, StringRef FS,
                                    const TargetOptions &Options,
                                    Optional<Reloc::Model> RM,
public:
   void addIRPasses() override;
   bool addPreISel() override;
   bool addILPOpts() override;
   bool addInstSelector() override;
   void addMachineSSAOptimization() override;
   void addPreRegAlloc() override;
   void addPreSched2() override;
   void addPreEmitPass() override;
+  ScheduleDAGInstrs *
+  createMachineScheduler(MachineSchedContext *C) const override {
+    const PPCSubtarget &ST = C->MF->getSubtarget<PPCSubtarget>();
+    if (ST.usePPCPreRASchedStrategy())
+      return createPPCMachineScheduler(C);
+    return nullptr;
+  }
+  ScheduleDAGInstrs *
+  createPostMachineScheduler(MachineSchedContext *C) const override {
+    const PPCSubtarget &ST = C->MF->getSubtarget<PPCSubtarget>();
+    if (ST.usePPCPostRASchedStrategy())
+      return createPPCPostMachineScheduler(C);
+    return nullptr;
+  }
 };
 
 } // end anonymous namespace
 
 TargetPassConfig *PPCTargetMachine::createPassConfig(PassManagerBase &PM) {
   return new PPCPassConfig(*this, PM);
 }
 
void PPCPassConfig::addPreEmitPass() {
   // Must run branch selection immediately preceding the asm printer.
   addPass(createPPCBranchSelectionPass(), false);
 }
 
 TargetTransformInfo
 PPCTargetMachine::getTargetTransformInfo(const Function &F) {
   return TargetTransformInfo(PPCTTIImpl(this, F));
 }
+
+static MachineSchedRegistry
+PPCPreRASchedRegistry("ppc-prera",
+                      "Run PowerPC PreRA specific scheduler",
+                      createPPCMachineScheduler);
+
+static MachineSchedRegistry
+PPCPostRASchedRegistry("ppc-postra",
+                       "Run PowerPC PostRA specific scheduler",
+                       createPPCPostMachineScheduler);