[EVM] [DO NOT MERGE] Add Ethereum stackification

llvm/lib/Target/EVM/CMakeLists.txt

@@ -20,20 +20,27 @@ add_llvm_target(EVMCodeGen
   EVMAllocaHoisting.cpp
   EVMArgumentMove.cpp
   EVMAsmPrinter.cpp
+  EVMBackwardPropagationStackification.cpp
   EVMCodegenPrepare.cpp
   EVMFrameLowering.cpp
   EVMISelDAGToDAG.cpp
   EVMISelLowering.cpp
   EVMInstrInfo.cpp
   EVMLinkRuntime.cpp
   EVMLowerIntrinsics.cpp
+  EVMMachineCFGInfo.cpp
   EVMMachineFunctionInfo.cpp
   EVMMCInstLower.cpp
   EVMOptimizeLiveIntervals.cpp
   EVMRegColoring.cpp
   EVMRegisterInfo.cpp
   EVMSingleUseExpression.cpp
+  EVMSplitCriticalEdges.cpp
+  EVMStackDebug.cpp
+  EVMStackLayoutGenerator.cpp
+  EVMStackModel.cpp
   EVMStackify.cpp
+  EVMStackifyCodeEmitter.cpp
   EVMSubtarget.cpp
   EVMTargetMachine.cpp
   EVMTargetTransformInfo.cpp

llvm/lib/Target/EVM/EVM.h

@@ -50,7 +50,9 @@ ModulePass *createEVMLinkRuntimePass();
 FunctionPass *createEVMOptimizeLiveIntervals();
 FunctionPass *createEVMRegColoring();
 FunctionPass *createEVMSingleUseExpression();
+FunctionPass *createEVMSplitCriticalEdges();
 FunctionPass *createEVMStackify();
+FunctionPass *createEVMBPStackification();
 
 // PassRegistry initialization declarations.
 void initializeEVMCodegenPreparePass(PassRegistry &);
@@ -61,7 +63,9 @@ void initializeEVMLinkRuntimePass(PassRegistry &);
 void initializeEVMOptimizeLiveIntervalsPass(PassRegistry &);
 void initializeEVMRegColoringPass(PassRegistry &);
 void initializeEVMSingleUseExpressionPass(PassRegistry &);
+void initializeEVMSplitCriticalEdgesPass(PassRegistry &);
 void initializeEVMStackifyPass(PassRegistry &);
+void initializeEVMBPStackificationPass(PassRegistry &);
 
 struct EVMLinkRuntimePass : PassInfoMixin<EVMLinkRuntimePass> {
   EVMLinkRuntimePass() = default;

llvm/lib/Target/EVM/EVMArgumentMove.cpp

@@ -6,7 +6,8 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// This file moves ARGUMENT instructions after ScheduleDAG scheduling.
+// This file moves and orders ARGUMENT instructions after ScheduleDAG
+// scheduling.
 //
 // Arguments are really live-in registers, however, since we use virtual
 // registers and LLVM doesn't support live-in virtual registers, we're
@@ -67,21 +68,24 @@ bool EVMArgumentMove::runOnMachineFunction(MachineFunction &MF) {
 
   bool Changed = false;
   MachineBasicBlock &EntryMBB = MF.front();
+  SmallVector<MachineInstr *> Args;
+  for (MachineInstr &MI : EntryMBB) {
+    if (EVM::ARGUMENT == MI.getOpcode())
+      Args.push_back(&MI);
+  }
 
-  // Look for the first NonArg instruction.
-  const auto InsertPt =
-      std::find_if_not(EntryMBB.begin(), EntryMBB.end(), [](auto &MI) {
-        return EVM::ARGUMENT == MI.getOpcode();
-      });
+  // Sort ARGUMENT instructions in ascending order of their arguments.
+  std::sort(Args.begin(), Args.end(),
+            [](const MachineInstr *MI1, const MachineInstr *MI2) {
+              int64_t Arg1Idx = MI1->getOperand(1).getImm();
+              int64_t Arg2Idx = MI2->getOperand(1).getImm();
+              return Arg1Idx < Arg2Idx;
+            });
 
-  // Now move any argument instructions later in the block
-  // to before our first NonArg instruction.
-  for (MachineInstr &MI : llvm::make_range(InsertPt, EntryMBB.end())) {
-    if (EVM::ARGUMENT == MI.getOpcode()) {
-      EntryMBB.insert(InsertPt, MI.removeFromParent());
-      Changed = true;
-    }
+  for (MachineInstr *MI : reverse(Args)) {
+    MachineInstr *Arg = MI->removeFromParent();
+    EntryMBB.insert(EntryMBB.begin(), Arg);
+    Changed = true;
   }
-
   return Changed;
 }

llvm/lib/Target/EVM/EVMAsmPrinter.cpp

@@ -31,6 +31,8 @@
 
 using namespace llvm;
 
+extern cl::opt<bool> EVMKeepRegisters;
+
 #define DEBUG_TYPE "asm-printer"
 
 namespace {
@@ -52,43 +54,22 @@ class EVMAsmPrinter : public AsmPrinter {
 
   StringRef getPassName() const override { return "EVM Assembly "; }
 
-  void SetupMachineFunction(MachineFunction &MF) override;
-
   void emitInstruction(const MachineInstr *MI) override;
 
-  void emitFunctionEntryLabel() override;
+  void emitBasicBlockStart(const MachineBasicBlock &MBB) override;
 
-  /// Return true if the basic block has exactly one predecessor and the control
-  /// transfer mechanism between the predecessor and this block is a
-  /// fall-through.
-  bool isBlockOnlyReachableByFallthrough(
-      const MachineBasicBlock *MBB) const override;
+  void emitFunctionEntryLabel() override;
 
   void emitEndOfAsmFile(Module &) override;
 
 private:
   void emitAssemblySymbol(const MachineInstr *MI);
   void emitWideRelocatableSymbol(const MachineInstr *MI);
   void emitLoadImmutableLabel(const MachineInstr *MI);
+  void emitJumpDest();
 };
 } // end of anonymous namespace
 
-void EVMAsmPrinter::SetupMachineFunction(MachineFunction &MF) {
-  // Unbundle <push_label, jump> bundles.
-  for (MachineBasicBlock &MBB : MF) {
-    MachineBasicBlock::instr_iterator I = MBB.instr_begin(),
-                                      E = MBB.instr_end();
-    for (; I != E; ++I) {
-      if (I->isBundledWithPred()) {
-        assert(I->isConditionalBranch() || I->isUnconditionalBranch());
-        I->unbundleFromPred();
-      }
-    }
-  }
-
-  AsmPrinter::SetupMachineFunction(MF);
-}
-
 void EVMAsmPrinter::emitFunctionEntryLabel() {
   AsmPrinter::emitFunctionEntryLabel();
 
@@ -111,19 +92,84 @@ void EVMAsmPrinter::emitFunctionEntryLabel() {
   }
 }
 
+void EVMAsmPrinter::emitBasicBlockStart(const MachineBasicBlock &MBB) {
+  AsmPrinter::emitBasicBlockStart(MBB);
+
+  // Emit JUMPDEST instruction at the beginning of the basic block, if
+  // this is not a block that is only reachable by fallthrough.
+  if (!EVMKeepRegisters && !AsmPrinter::isBlockOnlyReachableByFallthrough(&MBB))
+    emitJumpDest();
+}
+
 void EVMAsmPrinter::emitInstruction(const MachineInstr *MI) {
   EVMMCInstLower MCInstLowering(OutContext, *this, VRegMapping,
                                 MF->getRegInfo());
-  unsigned Opc = MI->getOpcode();
-  if (Opc == EVM::DATASIZE_S || Opc == EVM::DATAOFFSET_S) {
-    emitAssemblySymbol(MI);
+
+  switch (MI->getOpcode()) {
+  default:
+    break;
+  case EVM::PseudoCALL: {
+    // Generate push instruction with the address of a function.
+    MCInst Push;
+    Push.setOpcode(EVM::PUSH4_S);
+    assert(MI->getOperand(0).isGlobal() &&
+           "The first operand of PseudoCALL should be a GlobalValue.");
+
+    // TODO: #745: Refactor EVMMCInstLower::Lower so we could use lowerOperand
+    // instead of creating a MCOperand directly.
+    MCOperand MCOp = MCOperand::createExpr(MCSymbolRefExpr::create(
+        getSymbol(MI->getOperand(0).getGlobal()), OutContext));
+    Push.addOperand(MCOp);
+    EmitToStreamer(*OutStreamer, Push);
+
+    // Jump to a function.
+    MCInst Jump;
+    Jump.setOpcode(EVM::JUMP_S);
+    EmitToStreamer(*OutStreamer, Jump);
+
+    // In case a function has a return label, emit it, and also
+    // emit a JUMPDEST instruction.
+    if (MI->getNumExplicitOperands() > 1) {
+      assert(MI->getOperand(1).isMCSymbol() &&
+             "The second operand of PseudoCALL should be a MCSymbol.");
+      OutStreamer->emitLabel(MI->getOperand(1).getMCSymbol());
+      emitJumpDest();
+    }
     return;
   }
-  if (Opc == EVM::LINKERSYMBOL_S) {
-    emitWideRelocatableSymbol(MI);
+  case EVM::PseudoRET: {
+    // TODO: #746: Use PseudoInstExpansion and do this expansion in tblgen.
+    MCInst Jump;
+    Jump.setOpcode(EVM::JUMP_S);
+    EmitToStreamer(*OutStreamer, Jump);
     return;
   }
-  if (Opc == EVM::LOADIMMUTABLE_S) {
+  case EVM::PseudoJUMP:
+  case EVM::PseudoJUMPI: {
+    MCInst Push;
+    Push.setOpcode(EVM::PUSH4_S);
+
+    // TODO: #745: Refactor EVMMCInstLower::Lower so we could use lowerOperand
+    // instead of creating a MCOperand directly.
+    MCOperand MCOp = MCOperand::createExpr(MCSymbolRefExpr::create(
+        MI->getOperand(0).getMBB()->getSymbol(), OutContext));
+    Push.addOperand(MCOp);
+    EmitToStreamer(*OutStreamer, Push);
+
+    MCInst Jump;
+    Jump.setOpcode(MI->getOpcode() == EVM::PseudoJUMP ? EVM::JUMP_S
+                                                      : EVM::JUMPI_S);
+    EmitToStreamer(*OutStreamer, Jump);
+    return;
+  }
+  case EVM::LINKERSYMBOL_S:
+    emitWideRelocatableSymbol(MI);
+    return;
+  case EVM::DATASIZE_S:
+  case EVM::DATAOFFSET_S:
+    emitAssemblySymbol(MI);
+    return;
+  case EVM::LOADIMMUTABLE_S:
     emitLoadImmutableLabel(MI);
     return;
   }
@@ -133,12 +179,6 @@ void EVMAsmPrinter::emitInstruction(const MachineInstr *MI) {
   EmitToStreamer(*OutStreamer, TmpInst);
 }
 
-bool EVMAsmPrinter::isBlockOnlyReachableByFallthrough(
-    const MachineBasicBlock *MBB) const {
-  // For simplicity, always emit BB labels.
-  return false;
-}
-
 // Lowers LOADIMMUTABLE_S as show below:
 //   LOADIMMUTABLE_S @immutable_id
 //    ->
@@ -247,6 +287,12 @@ void EVMAsmPrinter::emitEndOfAsmFile(Module &) {
   ImmutablesMap.clear();
 }
 
+void EVMAsmPrinter::emitJumpDest() {
+  MCInst JumpDest;
+  JumpDest.setOpcode(EVM::JUMPDEST_S);
+  EmitToStreamer(*OutStreamer, JumpDest);
+}
+
 extern "C" LLVM_EXTERNAL_VISIBILITY void LLVMInitializeEVMAsmPrinter() {
   const RegisterAsmPrinter<EVMAsmPrinter> X(getTheEVMTarget());
 }

llvm/lib/Target/EVM/EVMBackwardPropagationStackification.cpp

@@ -0,0 +1,100 @@
+//===----- EVMBPStackification.cpp - BP stackification ---------*- 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
+//
+//===----------------------------------------------------------------------===//
+//
+// This file implements backward propagation (BP) stackification.
+// Original idea was taken from the Ethereum's compiler (solc) stackification
+// algorithm.
+// The algorithm is broken into following components:
+//   - CFG (Control Flow Graph) and CFG builder. Stackification CFG has similar
+//     structure to LLVM CFG one, but employs wider notion of instruction.
+//   - Stack layout generator. Contains information about the stack layout at
+//     entry and exit of each CFG::BasicBlock. It also contains input/output
+//     stack layout for each operation.
+//   - Code transformation into stakified form. This component uses both CFG
+//     and the stack layout information to get stackified LLVM MIR.
+//   - Stack shuffler. Finds optimal (locally) transformation between two stack
+//     layouts using three primitives: POP, PUSHn, DUPn. The stack shuffler
+//     is used by the components above.
+//
+//===----------------------------------------------------------------------===//
+
+#include "EVM.h"
+#include "EVMStackifyCodeEmitter.h"
+#include "EVMSubtarget.h"
+#include "llvm/CodeGen/LiveIntervals.h"
+#include "llvm/CodeGen/MachineLoopInfo.h"
+#include "llvm/CodeGen/Passes.h"
+#include "llvm/CodeGen/TargetInstrInfo.h"
+#include "llvm/InitializePasses.h"
+#include "llvm/Support/Debug.h"
+#include "llvm/Support/raw_ostream.h"
+
+using namespace llvm;
+
+#define DEBUG_TYPE "evm-ethereum-stackify"
+
+namespace {
+class EVMBPStackification final : public MachineFunctionPass {
+public:
+  static char ID; // Pass identification, replacement for typeid
+
+  EVMBPStackification() : MachineFunctionPass(ID) {}
+
+private:
+  StringRef getPassName() const override {
+    return "EVM Ethereum stackification";
+  }
+
+  void getAnalysisUsage(AnalysisUsage &AU) const override {
+    AU.addRequired<LiveIntervals>();
+    AU.addRequired<MachineLoopInfo>();
+    MachineFunctionPass::getAnalysisUsage(AU);
+  }
+
+  bool runOnMachineFunction(MachineFunction &MF) override;
+
+  MachineFunctionProperties getRequiredProperties() const override {
+    return MachineFunctionProperties().set(
+        MachineFunctionProperties::Property::TracksLiveness);
+  }
+};
+} // end anonymous namespace
+
+char EVMBPStackification::ID = 0;
+
+INITIALIZE_PASS_BEGIN(EVMBPStackification, DEBUG_TYPE,
+                      "Backward propagation stackification", false, false)
+INITIALIZE_PASS_DEPENDENCY(MachineLoopInfo)
+INITIALIZE_PASS_END(EVMBPStackification, DEBUG_TYPE,
+                    "Backward propagation stackification", false, false)
+
+FunctionPass *llvm::createEVMBPStackification() {
+  return new EVMBPStackification();
+}
+
+bool EVMBPStackification::runOnMachineFunction(MachineFunction &MF) {
+  LLVM_DEBUG({
+    dbgs() << "********** Backward propagation stackification **********\n"
+           << "********** Function: " << MF.getName() << '\n';
+  });
+
+  MachineRegisterInfo &MRI = MF.getRegInfo();
+  auto &LIS = getAnalysis<LiveIntervals>();
+  MachineLoopInfo *MLI = &getAnalysis<MachineLoopInfo>();
+
+  // We don't preserve SSA form.
+  MRI.leaveSSA();
+
+  assert(MRI.tracksLiveness() && "Stackification expects liveness");
+  EVMMachineCFGInfo CFGInfo(MF, MLI);
+  EVMStackModel StackModel(MF, LIS);
+  std::unique_ptr<EVMStackLayout> Layout =
+      EVMStackLayoutGenerator(MF, MLI, StackModel, CFGInfo).run();
+  EVMStackifyCodeEmitter(*Layout, StackModel, CFGInfo, MF).run();
+  return true;
+}

llvm/lib/Target/EVM/EVMInstrFormats.td

@@ -53,6 +53,7 @@ class NI<dag oops, dag iops, list<dag> pattern, bit stack,
   let Opc            = inst;
   let Inst{7-0}      = Opc;
   let GasCost        = cost;
+  let Defs           = !if(stack, [], [ARGUMENTS]);
 }
 
 // Generates both register and stack based versions of one actual instruction.
@@ -61,7 +62,7 @@ multiclass I<dag oops_r, dag iops_r, list<dag> pattern_r,
              int cost = 0, dag oops_s = (outs), dag iops_s = (ins), string argstr_s = ""> {
   let isCodeGenOnly = 1 in
   def "" : NI<oops_r, iops_r, pattern_r, false, opcstr#argstr_r, inst, cost>;
-  let BaseName = NAME in
+  let BaseName = NAME, Defs = []<Register> in
   def _S : NI<oops_s, iops_s, [], true, opcstr#argstr_s, inst, cost>;
 }
 
@@ -73,8 +74,8 @@ class NRI<dag oops, dag iops, list<dag> pattern, string asmstr>
   : NI<oops, iops, pattern, false, asmstr> {
 }
 
-class EVMPseudo<dag oops, dag iops, list<dag> pattern>
-  : NI<oops, iops, pattern, false> {
+class EVMPseudo<dag oops, dag iops, list<dag> pattern, bit stack = 0>
+  : NI<oops, iops, pattern, stack> {
   let isPseudo = 1;
   let isCodeGenOnly = 1;
 }