[llvm] r310044 - [mips][microMIPS] Extending size reduction pass with ADDIUSP and ADDIUR1SP

Zoran Jovanovic via llvm-commits llvm-commits at lists.llvm.org
Fri Aug 4 03:18:44 PDT 2017


Author: zjovanovic
Date: Fri Aug  4 03:18:44 2017
New Revision: 310044

URL: http://llvm.org/viewvc/llvm-project?rev=310044&view=rev
Log:
[mips][microMIPS] Extending size reduction pass with ADDIUSP and ADDIUR1SP
Author: milena.vujosevic.janicic
The patch extends size reduction pass for MicroMIPS.
The following instructions are examined and transformed, if possible:
ADDIU instruction is transformed into 16-bit instruction ADDIUSP
ADDIU instruction is transformed into 16-bit instruction ADDIUR1SP
Usage of u_int64_t replaced by uint64_t to avoid issues because of which previous patch version was reverted:
Differential Revision: https://reviews.llvm.org/D34511


Added:
    llvm/trunk/test/CodeGen/Mips/micromips-sizereduction/micromips-addiur1sp-addiusp.ll
Modified:
    llvm/trunk/lib/Target/Mips/MicroMipsSizeReduction.cpp

Modified: llvm/trunk/lib/Target/Mips/MicroMipsSizeReduction.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/Mips/MicroMipsSizeReduction.cpp?rev=310044&r1=310043&r2=310044&view=diff
==============================================================================
--- llvm/trunk/lib/Target/Mips/MicroMipsSizeReduction.cpp (original)
+++ llvm/trunk/lib/Target/Mips/MicroMipsSizeReduction.cpp Fri Aug  4 03:18:44 2017
@@ -32,6 +32,8 @@ namespace {
 enum OperandTransfer {
   OT_NA,          ///< Not applicable
   OT_OperandsAll, ///< Transfer all operands
+  OT_Operands02,  ///< Transfer operands 0 and 2
+  OT_Operand2,    ///< Transfer just operand 2
 };
 
 /// Reduction type
@@ -143,14 +145,23 @@ private:
   // returns true on success.
   static bool ReduceSXtoSX16(MachineInstr *MI, const ReduceEntry &Entry);
 
-  // Attempts to reduce arithmetic instructions, returns true on success
+  // Attempts to reduce arithmetic instructions, returns true on success.
   static bool ReduceArithmeticInstructions(MachineInstr *MI,
                                            const ReduceEntry &Entry);
 
-  // Changes opcode of an instruction
+  // Attempts to reduce ADDIU into ADDIUSP instruction,
+  // returns true on success.
+  static bool ReduceADDIUToADDIUSP(MachineInstr *MI, const ReduceEntry &Entry);
+
+  // Attempts to reduce ADDIU into ADDIUR1SP instruction,
+  // returns true on success.
+  static bool ReduceADDIUToADDIUR1SP(MachineInstr *MI,
+                                     const ReduceEntry &Entry);
+
+  // Changes opcode of an instruction.
   static bool ReplaceInstruction(MachineInstr *MI, const ReduceEntry &Entry);
 
-  // Table with transformation rules for each instruction
+  // Table with transformation rules for each instruction.
   static llvm::SmallVector<ReduceEntry, 16> ReduceTable;
 };
 
@@ -158,12 +169,20 @@ char MicroMipsSizeReduce::ID = 0;
 const MipsInstrInfo *MicroMipsSizeReduce::MipsII;
 
 // This table must be sorted by WideOpc as a main criterion and
-// ReduceType as a sub-criterion (when wide opcodes are the same)
+// ReduceType as a sub-criterion (when wide opcodes are the same).
 llvm::SmallVector<ReduceEntry, 16> MicroMipsSizeReduce::ReduceTable = {
 
     // ReduceType, OpCodes, ReduceFunction,
     // OpInfo(TransferOperands),
     // ImmField(Shift, LBound, HBound, ImmFieldPosition)
+    {RT_OneInstr, OpCodes(Mips::ADDiu, Mips::ADDIUR1SP_MM),
+     ReduceADDIUToADDIUR1SP, OpInfo(OT_Operands02), ImmField(2, 0, 64, 2)},
+    {RT_OneInstr, OpCodes(Mips::ADDiu, Mips::ADDIUSP_MM), ReduceADDIUToADDIUSP,
+     OpInfo(OT_Operand2), ImmField(0, 0, 0, 2)},
+    {RT_OneInstr, OpCodes(Mips::ADDiu_MM, Mips::ADDIUR1SP_MM),
+     ReduceADDIUToADDIUR1SP, OpInfo(OT_Operands02), ImmField(2, 0, 64, 2)},
+    {RT_OneInstr, OpCodes(Mips::ADDiu_MM, Mips::ADDIUSP_MM),
+     ReduceADDIUToADDIUSP, OpInfo(OT_Operand2), ImmField(0, 0, 0, 2)},
     {RT_OneInstr, OpCodes(Mips::ADDu, Mips::ADDU16_MM),
      ReduceArithmeticInstructions, OpInfo(OT_OperandsAll),
      ImmField(0, 0, 0, -1)},
@@ -174,6 +193,8 @@ llvm::SmallVector<ReduceEntry, 16> Micro
      OpInfo(OT_OperandsAll), ImmField(0, -1, 15, 2)},
     {RT_OneInstr, OpCodes(Mips::LBu_MM, Mips::LBU16_MM), ReduceLXUtoLXU16,
      OpInfo(OT_OperandsAll), ImmField(0, -1, 15, 2)},
+    {RT_OneInstr, OpCodes(Mips::LEA_ADDiu, Mips::ADDIUR1SP_MM),
+     ReduceADDIUToADDIUR1SP, OpInfo(OT_Operands02), ImmField(2, 0, 64, 2)},
     {RT_OneInstr, OpCodes(Mips::LHu, Mips::LHU16_MM), ReduceLXUtoLXU16,
      OpInfo(OT_OperandsAll), ImmField(1, 0, 16, 2)},
     {RT_OneInstr, OpCodes(Mips::LHu_MM, Mips::LHU16_MM), ReduceLXUtoLXU16,
@@ -201,9 +222,9 @@ llvm::SmallVector<ReduceEntry, 16> Micro
     {RT_OneInstr, OpCodes(Mips::SW_MM, Mips::SWSP_MM), ReduceXWtoXWSP,
      OpInfo(OT_OperandsAll), ImmField(2, 0, 32, 2)},
 };
-}
+} // namespace
 
-// Returns true if the machine operand MO is register SP
+// Returns true if the machine operand MO is register SP.
 static bool IsSP(const MachineOperand &MO) {
   if (MO.isReg() && ((MO.getReg() == Mips::SP)))
     return true;
@@ -225,7 +246,7 @@ static bool isMMSourceRegister(const Mac
 }
 
 // Returns true if the operand Op is an immediate value
-// and writes the immediate value into variable Imm
+// and writes the immediate value into variable Imm.
 static bool GetImm(MachineInstr *MI, unsigned Op, int64_t &Imm) {
 
   if (!MI->getOperand(Op).isImm())
@@ -234,17 +255,27 @@ static bool GetImm(MachineInstr *MI, uns
   return true;
 }
 
+// Returns true if the value is a valid immediate for ADDIUSP.
+static bool AddiuspImmValue(int64_t Value) {
+  int64_t Value2 = Value >> 2;
+  if (((Value & (int64_t)maskTrailingZeros<uint64_t>(2)) == Value) &&
+      ((Value2 >= 2 && Value2 <= 257) || (Value2 >= -258 && Value2 <= -3)))
+    return true;
+  return false;
+}
+
 // Returns true if the variable Value has the number of least-significant zero
-// bits equal to Shift and if the shifted value is between the bounds
+// bits equal to Shift and if the shifted value is between the bounds.
 static bool InRange(int64_t Value, unsigned short Shift, int LBound,
                     int HBound) {
   int64_t Value2 = Value >> Shift;
-  if ((Value2 << Shift) == Value && (Value2 >= LBound) && (Value2 < HBound))
+  if (((Value & (int64_t)maskTrailingZeros<uint64_t>(Shift)) == Value) &&
+      (Value2 >= LBound) && (Value2 < HBound))
     return true;
   return false;
 }
 
-// Returns true if immediate operand is in range
+// Returns true if immediate operand is in range.
 static bool ImmInRange(MachineInstr *MI, const ReduceEntry &Entry) {
 
   int64_t offset;
@@ -310,6 +341,34 @@ bool MicroMipsSizeReduce::ReduceArithmet
   return ReplaceInstruction(MI, Entry);
 }
 
+bool MicroMipsSizeReduce::ReduceADDIUToADDIUR1SP(MachineInstr *MI,
+                                                 const ReduceEntry &Entry) {
+
+  if (!ImmInRange(MI, Entry))
+    return false;
+
+  if (!isMMThreeBitGPRegister(MI->getOperand(0)) || !IsSP(MI->getOperand(1)))
+    return false;
+
+  return ReplaceInstruction(MI, Entry);
+}
+
+bool MicroMipsSizeReduce::ReduceADDIUToADDIUSP(MachineInstr *MI,
+                                               const ReduceEntry &Entry) {
+
+  int64_t ImmValue;
+  if (!GetImm(MI, Entry.ImmField(), ImmValue))
+    return false;
+
+  if (!AddiuspImmValue(ImmValue))
+    return false;
+
+  if (!IsSP(MI->getOperand(0)) || !IsSP(MI->getOperand(1)))
+    return false;
+
+  return ReplaceInstruction(MI, Entry);
+}
+
 bool MicroMipsSizeReduce::ReduceLXUtoLXU16(MachineInstr *MI,
                                            const ReduceEntry &Entry) {
 
@@ -361,10 +420,37 @@ bool MicroMipsSizeReduce::ReduceMBB(Mach
 bool MicroMipsSizeReduce::ReplaceInstruction(MachineInstr *MI,
                                              const ReduceEntry &Entry) {
 
-  MI->setDesc(MipsII->get(Entry.NarrowOpc()));
-  DEBUG(dbgs() << "Converted into 16-bit: " << *MI);
+  enum OperandTransfer OpTransfer = Entry.TransferOperands();
+
+  DEBUG(dbgs() << "Converting 32-bit: " << *MI);
   ++NumReduced;
-  return true;
+
+  if (OpTransfer == OT_OperandsAll) {
+    MI->setDesc(MipsII->get(Entry.NarrowOpc()));
+    DEBUG(dbgs() << "       to 16-bit: " << *MI);
+    return true;
+  } else {
+    MachineBasicBlock &MBB = *MI->getParent();
+    const MCInstrDesc &NewMCID = MipsII->get(Entry.NarrowOpc());
+    DebugLoc dl = MI->getDebugLoc();
+    MachineInstrBuilder MIB = BuildMI(MBB, MI, dl, NewMCID);
+
+    if (OpTransfer == OT_Operand2)
+      MIB.add(MI->getOperand(2));
+    else if (OpTransfer == OT_Operands02) {
+      MIB.add(MI->getOperand(0));
+      MIB.add(MI->getOperand(2));
+    } else
+      llvm_unreachable("Unknown operand transfer!");
+
+    // Transfer MI flags.
+    MIB.setMIFlags(MI->getFlags());
+
+    DEBUG(dbgs() << "       to 16-bit: " << *MIB);
+    MBB.erase_instr(MI);
+    return true;
+  }
+  return false;
 }
 
 bool MicroMipsSizeReduce::runOnMachineFunction(MachineFunction &MF) {
@@ -373,7 +459,8 @@ bool MicroMipsSizeReduce::runOnMachineFu
 
   // TODO: Add support for other subtargets:
   // microMIPS32r6 and microMIPS64r6
-  if (!Subtarget->inMicroMipsMode() || !Subtarget->hasMips32r2())
+  if (!Subtarget->inMicroMipsMode() || !Subtarget->hasMips32r2() ||
+      Subtarget->hasMips32r6())
     return false;
 
   MipsII = static_cast<const MipsInstrInfo *>(Subtarget->getInstrInfo());

Added: llvm/trunk/test/CodeGen/Mips/micromips-sizereduction/micromips-addiur1sp-addiusp.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/Mips/micromips-sizereduction/micromips-addiur1sp-addiusp.ll?rev=310044&view=auto
==============================================================================
--- llvm/trunk/test/CodeGen/Mips/micromips-sizereduction/micromips-addiur1sp-addiusp.ll (added)
+++ llvm/trunk/test/CodeGen/Mips/micromips-sizereduction/micromips-addiur1sp-addiusp.ll Fri Aug  4 03:18:44 2017
@@ -0,0 +1,17 @@
+; RUN: llc -march=mipsel -mcpu=mips32r2 -mattr=+micromips -verify-machineinstrs < %s | FileCheck %s
+
+define i32 @f1() {
+entry:
+; CHECK-LABEL: f1:
+; CHECK: addiusp
+; CHECK: addiur1sp
+; CHECK: addiusp
+  %a = alloca [10 x i32], align 4
+  %index = getelementptr inbounds [10 x i32], [10 x i32]* %a, i32 0, i32 0
+  call void @init(i32* %index)
+  %0 = load i32, i32* %index, align 4
+  ret i32 %0
+}
+
+declare void @init(i32*)
+




More information about the llvm-commits mailing list