AMDGPU/GlobalISel: Add stub reg-bank aware combiner pass

[kerbowa]

LGTM

[arsenm]

With updated boilerplate a8ca0ec267050f9ded865a729d50c2c0eb078b7e

[uabelho]

Hi!

With gcc 7.4, the following warning pops up with this patch:

[32/115] Building CXX object lib/Target/AMDGPU/CMakeFiles/LLVMAMDGPUCodeGen.dir/AMDGPURegBankCombiner.cpp.o
In file included from ../lib/Target/AMDGPU/AMDGPURegBankCombiner.cpp:39:0:
lib/Target/AMDGPU/AMDGPUGenRegBankGICombiner.inc:65:6: warning: 'bool {anonymous}::AMDGPUGenRegBankCombinerHelper::isRuleDisabled(unsigned int) const' defined but not used [-Wunused-function]
 bool AMDGPUGenRegBankCombinerHelper::isRuleDisabled(unsigned RuleID) const {
      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

llvm/lib/Target/AMDGPU/AMDGPU.h

 FunctionPass *createR600ExpandSpecialInstrsPass();
 FunctionPass *createR600EmitClauseMarkers();
 FunctionPass *createR600ClauseMergePass();
 FunctionPass *createR600Packetizer();
 FunctionPass *createR600ControlFlowFinalizer();
 FunctionPass *createAMDGPUCFGStructurizerPass();
 FunctionPass *createR600ISelDag(TargetMachine *TM, CodeGenOpt::Level OptLevel);
 
+// GlobalISel passes
+FunctionPass *createAMDGPURegBankCombiner(bool IsOptNone);
+void initializeAMDGPURegBankCombinerPass(PassRegistry &);
+
 // SI Passes
 FunctionPass *createGCNDPPCombinePass();
 FunctionPass *createSIAnnotateControlFlowPass();
 FunctionPass *createSIFoldOperandsPass();
 FunctionPass *createSIPeepholeSDWAPass();
 FunctionPass *createSILowerI1CopiesPass();
 FunctionPass *createSIFixupVectorISelPass();
 FunctionPass *createSIAddIMGInitPass();

llvm/lib/Target/AMDGPU/AMDGPURegBankCombiner.cpp

+//=== lib/Target/AMDGPU/AMDGPURegBankCombiner.cpp -------------------------===//
+//
+// 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
+//
+//===----------------------------------------------------------------------===//
+//
+// This pass does combining of machine instructions at the generic MI level,
+// before the legalizer.
+//
+//===----------------------------------------------------------------------===//
+
+#include "AMDGPUTargetMachine.h"
+#include "llvm/CodeGen/GlobalISel/Combiner.h"
+#include "llvm/CodeGen/GlobalISel/CombinerHelper.h"
+#include "llvm/CodeGen/GlobalISel/CombinerInfo.h"
+#include "llvm/CodeGen/GlobalISel/GISelKnownBits.h"
+#include "llvm/CodeGen/GlobalISel/MIPatternMatch.h"
+#include "llvm/CodeGen/MachineDominators.h"
+#include "llvm/CodeGen/MachineFunctionPass.h"
+#include "llvm/CodeGen/TargetPassConfig.h"
+#include "llvm/Support/Debug.h"
+
+#define DEBUG_TYPE "amdgpu-regbank-combiner"
+
+using namespace llvm;
+using namespace MIPatternMatch;
+
+namespace {
+
+class AMDGPURegBankCombinerInfo : public CombinerInfo {
+  GISelKnownBits *KB;
+  MachineDominatorTree *MDT;
+
+public:
+  AMDGPURegBankCombinerInfo(bool EnableOpt, bool OptSize, bool MinSize,
+                            GISelKnownBits *KB, MachineDominatorTree *MDT)
+      : CombinerInfo(/*AllowIllegalOps*/ true, /*ShouldLegalizeIllegal*/ false,
+                     /*LegalizerInfo*/ nullptr, EnableOpt, OptSize, MinSize),
+        KB(KB), MDT(MDT) {}
+
+  virtual bool combine(GISelChangeObserver &Observer, MachineInstr &MI,
+                       MachineIRBuilder &B) const override;
+};
+
+bool AMDGPURegBankCombinerInfo::combine(GISelChangeObserver &Observer,
+                                        MachineInstr &MI,
+                                        MachineIRBuilder &B) const {
+  CombinerHelper Helper(Observer, B, KB, MDT);
+
+  switch (MI.getOpcode()) {
+  default:
+    break;
+  }
+
+  return false;
+}
+
+// Pass boilerplate
+// ================
+
+class AMDGPURegBankCombiner : public MachineFunctionPass {
+public:
+  static char ID;
+
+  AMDGPURegBankCombiner(bool IsOptNone = false);
+
+  bool runOnMachineFunction(MachineFunction &MF) override;
+
+  void getAnalysisUsage(AnalysisUsage &AU) const override;
+
+private:
+  bool IsOptNone;
+};
+} // end anonymous namespace
+
+void AMDGPURegBankCombiner::getAnalysisUsage(AnalysisUsage &AU) const {
+  AU.addRequired<TargetPassConfig>();
+  AU.setPreservesCFG();
+  getSelectionDAGFallbackAnalysisUsage(AU);
+  AU.addRequired<GISelKnownBitsAnalysis>();
+  AU.addPreserved<GISelKnownBitsAnalysis>();
+  if (!IsOptNone) {
+    AU.addRequired<MachineDominatorTree>();
+    AU.addPreserved<MachineDominatorTree>();
+  }
+  MachineFunctionPass::getAnalysisUsage(AU);
+}
+
+AMDGPURegBankCombiner::AMDGPURegBankCombiner(bool IsOptNone)
+    : MachineFunctionPass(ID), IsOptNone(IsOptNone) {
+  initializeAMDGPURegBankCombinerPass(*PassRegistry::getPassRegistry());
+}
+
+bool AMDGPURegBankCombiner::runOnMachineFunction(MachineFunction &MF) {
+  if (MF.getProperties().hasProperty(
+          MachineFunctionProperties::Property::FailedISel))
+    return false;
+  auto *TPC = &getAnalysis<TargetPassConfig>();
+  const Function &F = MF.getFunction();
+  bool EnableOpt =
+      MF.getTarget().getOptLevel() != CodeGenOpt::None && !skipFunction(F);
+  GISelKnownBits *KB = &getAnalysis<GISelKnownBitsAnalysis>().get(MF);
+  MachineDominatorTree *MDT =
+      IsOptNone ? nullptr : &getAnalysis<MachineDominatorTree>();
+  AMDGPURegBankCombinerInfo PCInfo(EnableOpt, F.hasOptSize(), F.hasMinSize(),
+                                   KB, MDT);
+  Combiner C(PCInfo, TPC);
+  return C.combineMachineInstrs(MF, /*CSEInfo*/ nullptr);
+}
+
+char AMDGPURegBankCombiner::ID = 0;
+INITIALIZE_PASS_BEGIN(AMDGPURegBankCombiner, DEBUG_TYPE,
+                      "Combine AMDGPU machine instrs after regbankselect",
+                      false, false)
+INITIALIZE_PASS_DEPENDENCY(TargetPassConfig)
+INITIALIZE_PASS_DEPENDENCY(GISelKnownBitsAnalysis)
+INITIALIZE_PASS_END(AMDGPURegBankCombiner, DEBUG_TYPE,
+                    "Combine AMDGPU machine instrs after regbankselect", false,
+                    false)
+
+namespace llvm {
+FunctionPass *createAMDGPURegBankCombiner(bool IsOptNone) {
+  return new AMDGPURegBankCombiner(IsOptNone);
+}
+} // end namespace llvm

llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp

 
 bool GCNPassConfig::addLegalizeMachineIR() {
   addPass(new Legalizer());
   return false;
 }
 
 bool GCNPassConfig::addRegBankSelect() {
   addPass(new RegBankSelect());
+
+  bool IsOptNone = getOptLevel() == CodeGenOpt::None;
+  addPass(createAMDGPURegBankCombiner(IsOptNone));
+
   return false;
 }
 
 bool GCNPassConfig::addGlobalInstructionSelect() {
   addPass(new InstructionSelect());
   return false;
 }
 

llvm/lib/Target/AMDGPU/CMakeLists.txt

add_llvm_target(AMDGPUCodeGen
   AMDGPUMachineCFGStructurizer.cpp
   AMDGPUMachineFunction.cpp
   AMDGPUMachineModuleInfo.cpp
   AMDGPUMacroFusion.cpp
   AMDGPUMCInstLower.cpp
   AMDGPUOpenCLEnqueuedBlockLowering.cpp
   AMDGPUPromoteAlloca.cpp
   AMDGPUPropagateAttributes.cpp
+  AMDGPURegBankCombiner.cpp
   AMDGPURegisterBankInfo.cpp
   AMDGPURegisterInfo.cpp
   AMDGPURewriteOutArguments.cpp
   AMDGPUSubtarget.cpp
   AMDGPUTargetMachine.cpp
   AMDGPUTargetObjectFile.cpp
   AMDGPUTargetTransformInfo.cpp
   AMDGPUUnifyDivergentExitNodes.cpp