fpc/compiler/x86/aoptx86.pas

{
    Copyright (c) 1998-2002 by Florian Klaempfl and Jonas Maebe

    This unit contains the peephole optimizer.

    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 2 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program; if not, write to the Free Software
    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.

 ****************************************************************************
}
unit aoptx86;

{$i fpcdefs.inc}

{$define DEBUG_AOPTCPU}

  interface

    uses
      globtype,
      cpubase,
      aasmtai,aasmcpu,
      cgbase,cgutils,
      aopt,aoptobj;

    type
      TOptsToCheck = (
        aoc_MovAnd2Mov_3
      );

      TX86AsmOptimizer = class(TAsmOptimizer)
        { some optimizations are very expensive to check, so the
          pre opt pass can be used to set some flags, depending on the found
          instructions if it is worth to check a certain optimization }
        OptsToCheck : set of TOptsToCheck;
        function RegLoadedWithNewValue(reg : tregister; hp : tai) : boolean; override;
        function InstructionLoadsFromReg(const reg : TRegister; const hp : tai) : boolean; override;
        function RegReadByInstruction(reg : TRegister; hp : tai) : boolean;
        function RegInInstruction(Reg: TRegister; p1: tai): Boolean;override;
        function GetNextInstructionUsingReg(Current: tai; out Next: tai; reg: TRegister): Boolean;
        {
          In comparison with GetNextInstructionUsingReg, GetNextInstructionUsingRegTrackingUse tracks
          the use of a register by allocs/dealloc, so it can ignore calls.

          In the following example, GetNextInstructionUsingReg will return the second movq,
          GetNextInstructionUsingRegTrackingUse won't.

          movq	%rdi,%rax
          # Register rdi released

          # Register rdi allocated
          movq	%rax,%rdi

          While in this example:

          movq	%rdi,%rax
          call  proc
          movq	%rdi,%rax

          GetNextInstructionUsingRegTrackingUse will return the second instruction while GetNextInstructionUsingReg
          won't.
        }
        function GetNextInstructionUsingRegTrackingUse(Current: tai; out Next: tai; reg: TRegister): Boolean;
        function RegModifiedByInstruction(Reg: TRegister; p1: tai): boolean; override;
      private
        function SkipSimpleInstructions(var hp1: tai): Boolean;
      protected
        class function IsMOVZXAcceptable: Boolean; static; inline;

        { checks whether loading a new value in reg1 overwrites the entirety of reg2 }
        function Reg1WriteOverwritesReg2Entirely(reg1, reg2: tregister): boolean;
        { checks whether reading the value in reg1 depends on the value of reg2. This
          is very similar to SuperRegisterEquals, except it takes into account that
          R_SUBH and R_SUBL are independendent (e.g. reading from AL does not
          depend on the value in AH). }
        function Reg1ReadDependsOnReg2(reg1, reg2: tregister): boolean;

        { Replaces all references to AOldReg in a memory reference to ANewReg }
        class function ReplaceRegisterInRef(var ref: TReference; const AOldReg, ANewReg: TRegister): Boolean; static;

        { Replaces all references to AOldReg in an operand to ANewReg }
        class function ReplaceRegisterInOper(const p: taicpu; const OperIdx: Integer; const AOldReg, ANewReg: TRegister): Boolean; static;

        { Replaces all references to AOldReg in an instruction to ANewReg,
          except where the register is being written }
        function ReplaceRegisterInInstruction(const p: taicpu; const AOldReg, ANewReg: TRegister): Boolean;

        { Returns true if the reference only refers to ESP or EBP (or their 64-bit equivalents),
          or writes to a global symbol }
        class function IsRefSafe(const ref: PReference): Boolean; static; inline;


        { Returns true if the given MOV instruction can be safely converted to CMOV }
        class function CanBeCMOV(p : tai) : boolean; static;


        { Converts the LEA instruction to ADD/INC/SUB/DEC. Returns True if the
          conversion was successful }
        function ConvertLEA(const p : taicpu): Boolean;

        function DeepMOVOpt(const p_mov: taicpu; const hp: taicpu): Boolean;

        procedure DebugMsg(const s : string; p : tai);inline;

        class function IsExitCode(p : tai) : boolean; static;
        class function isFoldableArithOp(hp1 : taicpu; reg : tregister) : boolean; static;
        procedure RemoveLastDeallocForFuncRes(p : tai);

        function DoSubAddOpt(var p : tai) : Boolean;

        function PrePeepholeOptSxx(var p : tai) : boolean;
        function PrePeepholeOptIMUL(var p : tai) : boolean;

        function OptPass1AND(var p : tai) : boolean;
        function OptPass1_V_MOVAP(var p : tai) : boolean;
        function OptPass1VOP(var p : tai) : boolean;
        function OptPass1MOV(var p : tai) : boolean;
        function OptPass1Movx(var p : tai) : boolean;
        function OptPass1MOVXX(var p : tai) : boolean;
        function OptPass1OP(var p : tai) : boolean;
        function OptPass1LEA(var p : tai) : boolean;
        function OptPass1Sub(var p : tai) : boolean;
        function OptPass1SHLSAL(var p : tai) : boolean;
        function OptPass1SETcc(var p : tai) : boolean;
        function OptPass1FSTP(var p : tai) : boolean;
        function OptPass1FLD(var p : tai) : boolean;
        function OptPass1Cmp(var p : tai) : boolean;
        function OptPass1PXor(var p : tai) : boolean;
        function OptPass1VPXor(var p: tai): boolean;
        function OptPass1Imul(var p : tai) : boolean;

        function OptPass2MOV(var p : tai) : boolean;
        function OptPass2Imul(var p : tai) : boolean;
        function OptPass2Jmp(var p : tai) : boolean;
        function OptPass2Jcc(var p : tai) : boolean;
        function OptPass2Lea(var p: tai): Boolean;
        function OptPass2SUB(var p: tai): Boolean;

        function PostPeepholeOptMov(var p : tai) : Boolean;
{$ifdef x86_64} { These post-peephole optimisations only affect 64-bit registers. [Kit] }
        function PostPeepholeOptMovzx(var p : tai) : Boolean;
        function PostPeepholeOptXor(var p : tai) : Boolean;
{$endif}
        function PostPeepholeOptMOVSX(var p : tai) : boolean;
        function PostPeepholeOptCmp(var p : tai) : Boolean;
        function PostPeepholeOptTestOr(var p : tai) : Boolean;
        function PostPeepholeOptCall(var p : tai) : Boolean;
        function PostPeepholeOptLea(var p : tai) : Boolean;
        function PostPeepholeOptPush(var p: tai): Boolean;

        procedure ConvertJumpToRET(const p: tai; const ret_p: tai);

        { Processor-dependent reference optimisation }
        class procedure OptimizeRefs(var p: taicpu); static;
      end;

    function MatchInstruction(const instr: tai; const op: TAsmOp; const opsize: topsizes): boolean;
    function MatchInstruction(const instr: tai; const op1,op2: TAsmOp; const opsize: topsizes): boolean;
    function MatchInstruction(const instr: tai; const op1,op2,op3: TAsmOp; const opsize: topsizes): boolean;
    function MatchInstruction(const instr: tai; const ops: array of TAsmOp; const opsize: topsizes): boolean;

    function MatchOperand(const oper: TOper; const reg: TRegister): boolean; inline;
    function MatchOperand(const oper: TOper; const a: tcgint): boolean; inline;
    function MatchOperand(const oper1: TOper; const oper2: TOper): boolean;
{$if max_operands>2}
    function MatchOperand(const oper1: TOper; const oper2: TOper; const oper3: TOper): boolean;
{$endif max_operands>2}

    function RefsEqual(const r1, r2: treference): boolean;

    function MatchReference(const ref : treference;base,index : TRegister) : Boolean;

    { returns true, if ref is a reference using only the registers passed as base and index
      and having an offset }
    function MatchReferenceWithOffset(const ref : treference;base,index : TRegister) : Boolean;

  implementation

    uses
      cutils,verbose,
      systems,
      globals,
      cpuinfo,
      procinfo,
      paramgr,
      aasmbase,
      aoptbase,aoptutils,
      symconst,symsym,
      cgx86,
      itcpugas;

{$ifdef DEBUG_AOPTCPU}
    const
      SPeepholeOptimization: shortstring = 'Peephole Optimization: ';
{$else DEBUG_AOPTCPU}
    { Empty strings help the optimizer to remove string concatenations that won't
      ever appear to the user on release builds. [Kit] }
    const
      SPeepholeOptimization = '';
{$endif DEBUG_AOPTCPU}

    function MatchInstruction(const instr: tai; const op: TAsmOp; const opsize: topsizes): boolean;
      begin
        result :=
          (instr.typ = ait_instruction) and
          (taicpu(instr).opcode = op) and
          ((opsize = []) or (taicpu(instr).opsize in opsize));
      end;


    function MatchInstruction(const instr: tai; const op1,op2: TAsmOp; const opsize: topsizes): boolean;
      begin
        result :=
          (instr.typ = ait_instruction) and
          ((taicpu(instr).opcode = op1) or
           (taicpu(instr).opcode = op2)
          ) and
          ((opsize = []) or (taicpu(instr).opsize in opsize));
      end;


    function MatchInstruction(const instr: tai; const op1,op2,op3: TAsmOp; const opsize: topsizes): boolean;
      begin
        result :=
          (instr.typ = ait_instruction) and
          ((taicpu(instr).opcode = op1) or
           (taicpu(instr).opcode = op2) or
           (taicpu(instr).opcode = op3)
          ) and
          ((opsize = []) or (taicpu(instr).opsize in opsize));
      end;


    function MatchInstruction(const instr : tai;const ops : array of TAsmOp;
     const opsize : topsizes) : boolean;
      var
        op : TAsmOp;
      begin
        result:=false;
        for op in ops do
          begin
            if (instr.typ = ait_instruction) and
               (taicpu(instr).opcode = op) and
               ((opsize = []) or (taicpu(instr).opsize in opsize)) then
               begin
                 result:=true;
                 exit;
               end;
          end;
      end;


    function MatchOperand(const oper: TOper; const reg: TRegister): boolean; inline;
      begin
        result := (oper.typ = top_reg) and (oper.reg = reg);
      end;


    function MatchOperand(const oper: TOper; const a: tcgint): boolean; inline;
      begin
        result := (oper.typ = top_const) and (oper.val = a);
      end;


    function MatchOperand(const oper1: TOper; const oper2: TOper): boolean;
      begin
        result := oper1.typ = oper2.typ;

        if result then
          case oper1.typ of
            top_const:
              Result:=oper1.val = oper2.val;
            top_reg:
              Result:=oper1.reg = oper2.reg;
            top_ref:
              Result:=RefsEqual(oper1.ref^, oper2.ref^);
            else
              internalerror(2013102801);
          end
      end;


    function MatchOperand(const oper1: TOper; const oper2: TOper; const oper3: TOper): boolean;
      begin
        result := (oper1.typ = oper2.typ) and (oper1.typ = oper3.typ);

        if result then
          case oper1.typ of
            top_const:
              Result:=(oper1.val = oper2.val) and (oper1.val = oper3.val);
            top_reg:
              Result:=(oper1.reg = oper2.reg) and (oper1.reg = oper3.reg);
            top_ref:
              Result:=RefsEqual(oper1.ref^, oper2.ref^) and RefsEqual(oper1.ref^, oper3.ref^);
            else
              internalerror(2020052401);
          end
      end;


    function RefsEqual(const r1, r2: treference): boolean;
      begin
        RefsEqual :=
          (r1.offset = r2.offset) and
          (r1.segment = r2.segment) and (r1.base = r2.base) and
          (r1.index = r2.index) and (r1.scalefactor = r2.scalefactor) and
          (r1.symbol=r2.symbol) and (r1.refaddr = r2.refaddr) and
          (r1.relsymbol = r2.relsymbol) and
          (r1.volatility=[]) and
          (r2.volatility=[]);
      end;


    function MatchReference(const ref : treference;base,index : TRegister) : Boolean;
      begin
       Result:=(ref.offset=0) and
         (ref.scalefactor in [0,1]) and
         (ref.segment=NR_NO) and
         (ref.symbol=nil) and
         (ref.relsymbol=nil) and
         ((base=NR_INVALID) or
          (ref.base=base)) and
         ((index=NR_INVALID) or
          (ref.index=index)) and
         (ref.volatility=[]);
      end;


    function MatchReferenceWithOffset(const ref : treference;base,index : TRegister) : Boolean;
      begin
       Result:=(ref.scalefactor in [0,1]) and
         (ref.segment=NR_NO) and
         (ref.symbol=nil) and
         (ref.relsymbol=nil) and
         ((base=NR_INVALID) or
          (ref.base=base)) and
         ((index=NR_INVALID) or
          (ref.index=index)) and
         (ref.volatility=[]);
      end;


    function InstrReadsFlags(p: tai): boolean;
      begin
        InstrReadsFlags := true;
        case p.typ of
          ait_instruction:
            if InsProp[taicpu(p).opcode].Ch*
               [Ch_RCarryFlag,Ch_RParityFlag,Ch_RAuxiliaryFlag,Ch_RZeroFlag,Ch_RSignFlag,Ch_ROverflowFlag,
                Ch_RWCarryFlag,Ch_RWParityFlag,Ch_RWAuxiliaryFlag,Ch_RWZeroFlag,Ch_RWSignFlag,Ch_RWOverflowFlag,
                Ch_RFlags,Ch_RWFlags,Ch_RFLAGScc,Ch_All]<>[] then
              exit;
          ait_label:
            exit;
          else
            ;
        end;
        InstrReadsFlags := false;
      end;


  function TX86AsmOptimizer.GetNextInstructionUsingReg(Current: tai; out Next: tai; reg: TRegister): Boolean;
    begin
      Next:=Current;
      repeat
        Result:=GetNextInstruction(Next,Next);
      until not (Result) or
            not(cs_opt_level3 in current_settings.optimizerswitches) or
            (Next.typ<>ait_instruction) or
            RegInInstruction(reg,Next) or
            is_calljmp(taicpu(Next).opcode);
    end;


  function TX86AsmOptimizer.GetNextInstructionUsingRegTrackingUse(Current: tai; out Next: tai; reg: TRegister): Boolean;
    begin
      if not(cs_opt_level3 in current_settings.optimizerswitches) then
        begin
          Result:=GetNextInstruction(Current,Next);
          exit;
        end;
      Next:=tai(Current.Next);
      Result:=false;
      while assigned(Next) do
        begin
          if ((Next.typ=ait_instruction) and is_calljmp(taicpu(Next).opcode) and not(taicpu(Next).opcode=A_CALL)) or
            ((Next.typ=ait_regalloc) and (getsupreg(tai_regalloc(Next).reg)=getsupreg(reg))) or
            ((Next.typ=ait_label) and not(labelCanBeSkipped(Tai_Label(Next)))) then
            exit
          else if (Next.typ=ait_instruction) and RegInInstruction(reg,Next) and not(taicpu(Next).opcode=A_CALL) then
            begin
              Result:=true;
              exit;
            end;
          Next:=tai(Next.Next);
        end;
    end;


  function TX86AsmOptimizer.InstructionLoadsFromReg(const reg: TRegister;const hp: tai): boolean;
    begin
      Result:=RegReadByInstruction(reg,hp);
    end;


  function TX86AsmOptimizer.RegReadByInstruction(reg: TRegister; hp: tai): boolean;
    var
      p: taicpu;
      opcount: longint;
    begin
      RegReadByInstruction := false;
      if hp.typ <> ait_instruction then
        exit;
      p := taicpu(hp);
      case p.opcode of
        A_CALL:
          regreadbyinstruction := true;
        A_IMUL:
          case p.ops of
            1:
              regReadByInstruction := RegInOp(reg,p.oper[0]^) or
                 (
                  ((getregtype(reg)=R_INTREGISTER) and (getsupreg(reg)=RS_EAX)) and
                  ((getsubreg(reg)<>R_SUBH) or (p.opsize<>S_B))
                 );
            2,3:
              regReadByInstruction :=
                reginop(reg,p.oper[0]^) or
                reginop(reg,p.oper[1]^);
            else
              InternalError(2019112801);
          end;
        A_MUL:
          begin
            regReadByInstruction := RegInOp(reg,p.oper[0]^) or
               (
                ((getregtype(reg)=R_INTREGISTER) and (getsupreg(reg)=RS_EAX)) and
                ((getsubreg(reg)<>R_SUBH) or (p.opsize<>S_B))
               );
          end;
        A_IDIV,A_DIV:
          begin
            regReadByInstruction := RegInOp(reg,p.oper[0]^) or
               (
                 (getregtype(reg)=R_INTREGISTER) and
                 (
                   (getsupreg(reg)=RS_EAX) or ((getsupreg(reg)=RS_EDX) and (p.opsize<>S_B))
                 )
               );
          end;
        else
          begin
            if (p.opcode=A_LEA) and is_segment_reg(reg) then
              begin
                RegReadByInstruction := false;
                exit;
              end;
            for opcount := 0 to p.ops-1 do
              if (p.oper[opCount]^.typ = top_ref) and
                 RegInRef(reg,p.oper[opcount]^.ref^) then
                begin
                  RegReadByInstruction := true;
                  exit
                end;
            { special handling for SSE MOVSD }
            if (p.opcode=A_MOVSD) and (p.ops>0) then
              begin
                if p.ops<>2 then
                  internalerror(2017042702);
                regReadByInstruction := reginop(reg,p.oper[0]^) or
                  (
                   (p.oper[1]^.typ=top_reg) and (p.oper[0]^.typ=top_reg) and reginop(reg, p.oper[1]^)
                  );
                exit;
              end;
            with insprop[p.opcode] do
              begin
                if getregtype(reg)=R_INTREGISTER then
                  begin
                    case getsupreg(reg) of
                      RS_EAX:
                        if [Ch_REAX,Ch_RWEAX,Ch_MEAX]*Ch<>[] then
                          begin
                            RegReadByInstruction := true;
                            exit
                          end;
                      RS_ECX:
                        if [Ch_RECX,Ch_RWECX,Ch_MECX]*Ch<>[] then
                          begin
                            RegReadByInstruction := true;
                            exit
                          end;
                      RS_EDX:
                        if [Ch_REDX,Ch_RWEDX,Ch_MEDX]*Ch<>[] then
                          begin
                            RegReadByInstruction := true;
                            exit
                          end;
                      RS_EBX:
                        if [Ch_REBX,Ch_RWEBX,Ch_MEBX]*Ch<>[] then
                          begin
                            RegReadByInstruction := true;
                            exit
                          end;
                      RS_ESP:
                        if [Ch_RESP,Ch_RWESP,Ch_MESP]*Ch<>[] then
                          begin
                            RegReadByInstruction := true;
                            exit
                          end;
                      RS_EBP:
                        if [Ch_REBP,Ch_RWEBP,Ch_MEBP]*Ch<>[] then
                          begin
                            RegReadByInstruction := true;
                            exit
                          end;
                      RS_ESI:
                        if [Ch_RESI,Ch_RWESI,Ch_MESI]*Ch<>[] then
                          begin
                            RegReadByInstruction := true;
                            exit
                          end;
                      RS_EDI:
                        if [Ch_REDI,Ch_RWEDI,Ch_MEDI]*Ch<>[] then
                          begin
                            RegReadByInstruction := true;
                            exit
                          end;
                    end;
                  end;
                if SuperRegistersEqual(reg,NR_DEFAULTFLAGS) then
                  begin
                    if (Ch_RFLAGScc in Ch) and not(getsubreg(reg) in [R_SUBW,R_SUBD,R_SUBQ]) then
                      begin
                        case p.condition of
                          C_A,C_NBE,       { CF=0 and ZF=0  }
                          C_BE,C_NA:       { CF=1 or ZF=1   }
                            RegReadByInstruction:=getsubreg(reg) in [R_SUBFLAGCARRY,R_SUBFLAGZERO];
                          C_AE,C_NB,C_NC,  { CF=0           }
                          C_B,C_NAE,C_C:   { CF=1           }
                            RegReadByInstruction:=getsubreg(reg) in [R_SUBFLAGCARRY];
                          C_NE,C_NZ,       { ZF=0           }
                          C_E,C_Z:         { ZF=1           }
                            RegReadByInstruction:=getsubreg(reg) in [R_SUBFLAGZERO];
                          C_G,C_NLE,       { ZF=0 and SF=OF }
                          C_LE,C_NG:       { ZF=1 or SF<>OF }
                            RegReadByInstruction:=getsubreg(reg) in [R_SUBFLAGZERO,R_SUBFLAGSIGN,R_SUBFLAGOVERFLOW];
                          C_GE,C_NL,       { SF=OF          }
                          C_L,C_NGE:       { SF<>OF         }
                            RegReadByInstruction:=getsubreg(reg) in [R_SUBFLAGSIGN,R_SUBFLAGOVERFLOW];
                          C_NO,            { OF=0           }
                          C_O:             { OF=1           }
                            RegReadByInstruction:=getsubreg(reg) in [R_SUBFLAGOVERFLOW];
                          C_NP,C_PO,       { PF=0           }
                          C_P,C_PE:        { PF=1           }
                            RegReadByInstruction:=getsubreg(reg) in [R_SUBFLAGPARITY];
                          C_NS,            { SF=0           }
                          C_S:             { SF=1           }
                            RegReadByInstruction:=getsubreg(reg) in [R_SUBFLAGSIGN];
                          else
                            internalerror(2017042701);
                        end;
                        if RegReadByInstruction then
                          exit;
                      end;
                    case getsubreg(reg) of
                      R_SUBW,R_SUBD,R_SUBQ:
                        RegReadByInstruction :=
                          [Ch_RCarryFlag,Ch_RParityFlag,Ch_RAuxiliaryFlag,Ch_RZeroFlag,Ch_RSignFlag,Ch_ROverflowFlag,
                           Ch_RWCarryFlag,Ch_RWParityFlag,Ch_RWAuxiliaryFlag,Ch_RWZeroFlag,Ch_RWSignFlag,Ch_RWOverflowFlag,
                           Ch_RDirFlag,Ch_RFlags,Ch_RWFlags,Ch_RFLAGScc]*Ch<>[];
                      R_SUBFLAGCARRY:
                        RegReadByInstruction:=[Ch_RCarryFlag,Ch_RWCarryFlag,Ch_RFlags,Ch_RWFlags]*Ch<>[];
                      R_SUBFLAGPARITY:
                        RegReadByInstruction:=[Ch_RParityFlag,Ch_RWParityFlag,Ch_RFlags,Ch_RWFlags]*Ch<>[];
                      R_SUBFLAGAUXILIARY:
                        RegReadByInstruction:=[Ch_RAuxiliaryFlag,Ch_RWAuxiliaryFlag,Ch_RFlags,Ch_RWFlags]*Ch<>[];
                      R_SUBFLAGZERO:
                        RegReadByInstruction:=[Ch_RZeroFlag,Ch_RWZeroFlag,Ch_RFlags,Ch_RWFlags]*Ch<>[];
                      R_SUBFLAGSIGN:
                        RegReadByInstruction:=[Ch_RSignFlag,Ch_RWSignFlag,Ch_RFlags,Ch_RWFlags]*Ch<>[];
                      R_SUBFLAGOVERFLOW:
                        RegReadByInstruction:=[Ch_ROverflowFlag,Ch_RWOverflowFlag,Ch_RFlags,Ch_RWFlags]*Ch<>[];
                      R_SUBFLAGINTERRUPT:
                        RegReadByInstruction:=[Ch_RFlags,Ch_RWFlags]*Ch<>[];
                      R_SUBFLAGDIRECTION:
                        RegReadByInstruction:=[Ch_RDirFlag,Ch_RFlags,Ch_RWFlags]*Ch<>[];
                      else
                        internalerror(2017042601);
                    end;
                    exit;
                  end;
                if (Ch_NoReadIfEqualRegs in Ch) and (p.ops=2) and
                   (p.oper[0]^.typ=top_reg) and (p.oper[1]^.typ=top_reg) and
                   (p.oper[0]^.reg=p.oper[1]^.reg) then
                  exit;
                if ([CH_RWOP1,CH_ROP1,CH_MOP1]*Ch<>[]) and reginop(reg,p.oper[0]^) then
                  begin
                    RegReadByInstruction := true;
                    exit
                  end;
                if ([Ch_RWOP2,Ch_ROP2,Ch_MOP2]*Ch<>[]) and reginop(reg,p.oper[1]^) then
                  begin
                    RegReadByInstruction := true;
                    exit
                  end;
                if ([Ch_RWOP3,Ch_ROP3,Ch_MOP3]*Ch<>[]) and reginop(reg,p.oper[2]^) then
                  begin
                    RegReadByInstruction := true;
                    exit
                  end;
                if ([Ch_RWOP4,Ch_ROP4,Ch_MOP4]*Ch<>[]) and reginop(reg,p.oper[3]^) then
                  begin
                    RegReadByInstruction := true;
                    exit
                  end;
              end;
          end;
      end;
    end;


  function TX86AsmOptimizer.RegInInstruction(Reg: TRegister; p1: tai): Boolean;
    begin
      result:=false;
      if p1.typ<>ait_instruction then
        exit;

      if (Ch_All in insprop[taicpu(p1).opcode].Ch) then
        exit(true);

      if (getregtype(reg)=R_INTREGISTER) and
        { change information for xmm movsd are not correct }
        ((taicpu(p1).opcode<>A_MOVSD) or (taicpu(p1).ops=0)) then
        begin
          case getsupreg(reg) of
            { RS_EAX = RS_RAX on x86-64 }
            RS_EAX:
              result:=([Ch_REAX,Ch_RRAX,Ch_WEAX,Ch_WRAX,Ch_RWEAX,Ch_RWRAX,Ch_MEAX,Ch_MRAX]*insprop[taicpu(p1).opcode].Ch)<>[];
            RS_ECX:
              result:=([Ch_RECX,Ch_RRCX,Ch_WECX,Ch_WRCX,Ch_RWECX,Ch_RWRCX,Ch_MECX,Ch_MRCX]*insprop[taicpu(p1).opcode].Ch)<>[];
            RS_EDX:
              result:=([Ch_REDX,Ch_RRDX,Ch_WEDX,Ch_WRDX,Ch_RWEDX,Ch_RWRDX,Ch_MEDX,Ch_MRDX]*insprop[taicpu(p1).opcode].Ch)<>[];
            RS_EBX:
              result:=([Ch_REBX,Ch_RRBX,Ch_WEBX,Ch_WRBX,Ch_RWEBX,Ch_RWRBX,Ch_MEBX,Ch_MRBX]*insprop[taicpu(p1).opcode].Ch)<>[];
            RS_ESP:
              result:=([Ch_RESP,Ch_RRSP,Ch_WESP,Ch_WRSP,Ch_RWESP,Ch_RWRSP,Ch_MESP,Ch_MRSP]*insprop[taicpu(p1).opcode].Ch)<>[];
            RS_EBP:
              result:=([Ch_REBP,Ch_RRBP,Ch_WEBP,Ch_WRBP,Ch_RWEBP,Ch_RWRBP,Ch_MEBP,Ch_MRBP]*insprop[taicpu(p1).opcode].Ch)<>[];
            RS_ESI:
              result:=([Ch_RESI,Ch_RRSI,Ch_WESI,Ch_WRSI,Ch_RWESI,Ch_RWRSI,Ch_MESI,Ch_MRSI,Ch_RMemEDI]*insprop[taicpu(p1).opcode].Ch)<>[];
            RS_EDI:
              result:=([Ch_REDI,Ch_RRDI,Ch_WEDI,Ch_WRDI,Ch_RWEDI,Ch_RWRDI,Ch_MEDI,Ch_MRDI,Ch_WMemEDI]*insprop[taicpu(p1).opcode].Ch)<>[];
            else
              ;
          end;
          if result then
            exit;
        end
      else if SuperRegistersEqual(reg,NR_DEFAULTFLAGS) then
        begin
          if ([Ch_RFlags,Ch_WFlags,Ch_RWFlags,Ch_RFLAGScc]*insprop[taicpu(p1).opcode].Ch)<>[] then
            exit(true);
          case getsubreg(reg) of
            R_SUBFLAGCARRY:
              Result:=([Ch_RCarryFlag,Ch_RWCarryFlag,Ch_W0CarryFlag,Ch_W1CarryFlag,Ch_WCarryFlag,Ch_WUCarryFlag]*insprop[taicpu(p1).opcode].Ch)<>[];
            R_SUBFLAGPARITY:
              Result:=([Ch_RParityFlag,Ch_RWParityFlag,Ch_W0ParityFlag,Ch_W1ParityFlag,Ch_WParityFlag,Ch_WUParityFlag]*insprop[taicpu(p1).opcode].Ch)<>[];
            R_SUBFLAGAUXILIARY:
              Result:=([Ch_RAuxiliaryFlag,Ch_RWAuxiliaryFlag,Ch_W0AuxiliaryFlag,Ch_W1AuxiliaryFlag,Ch_WAuxiliaryFlag,Ch_WUAuxiliaryFlag]*insprop[taicpu(p1).opcode].Ch)<>[];
            R_SUBFLAGZERO:
              Result:=([Ch_RZeroFlag,Ch_RWZeroFlag,Ch_W0ZeroFlag,Ch_W1ZeroFlag,Ch_WZeroFlag,Ch_WUZeroFlag]*insprop[taicpu(p1).opcode].Ch)<>[];
            R_SUBFLAGSIGN:
              Result:=([Ch_RSignFlag,Ch_RWSignFlag,Ch_W0SignFlag,Ch_W1SignFlag,Ch_WSignFlag,Ch_WUSignFlag]*insprop[taicpu(p1).opcode].Ch)<>[];
            R_SUBFLAGOVERFLOW:
              Result:=([Ch_ROverflowFlag,Ch_RWOverflowFlag,Ch_W0OverflowFlag,Ch_W1OverflowFlag,Ch_WOverflowFlag,Ch_WUOverflowFlag]*insprop[taicpu(p1).opcode].Ch)<>[];
            R_SUBFLAGINTERRUPT:
              Result:=([Ch_W0IntFlag,Ch_W1IntFlag,Ch_WFlags]*insprop[taicpu(p1).opcode].Ch)<>[];
            R_SUBFLAGDIRECTION:
              Result:=([Ch_RDirFlag,Ch_W0DirFlag,Ch_W1DirFlag,Ch_WFlags]*insprop[taicpu(p1).opcode].Ch)<>[];
            else
              ;
          end;
          if result then
            exit;
        end
      else if (getregtype(reg)=R_FPUREGISTER) and (Ch_FPU in insprop[taicpu(p1).opcode].Ch) then
        exit(true);
      Result:=inherited RegInInstruction(Reg, p1);
    end;


    function TX86AsmOptimizer.RegModifiedByInstruction(Reg: TRegister; p1: tai): boolean;
      begin
        Result := False;
        if p1.typ <> ait_instruction then
          exit;

        with insprop[taicpu(p1).opcode] do
          if SuperRegistersEqual(reg,NR_DEFAULTFLAGS) then
            begin
              case getsubreg(reg) of
                R_SUBW,R_SUBD,R_SUBQ:
                  Result :=
                    [Ch_WCarryFlag,Ch_WParityFlag,Ch_WAuxiliaryFlag,Ch_WZeroFlag,Ch_WSignFlag,Ch_WOverflowFlag,
                     Ch_RWCarryFlag,Ch_RWParityFlag,Ch_RWAuxiliaryFlag,Ch_RWZeroFlag,Ch_RWSignFlag,Ch_RWOverflowFlag,
                     Ch_W0DirFlag,Ch_W1DirFlag,Ch_W0IntFlag,Ch_W1IntFlag,Ch_WFlags,Ch_RWFlags]*Ch<>[];
                R_SUBFLAGCARRY:
                  Result:=[Ch_WCarryFlag,Ch_RWCarryFlag,Ch_WFlags,Ch_RWFlags]*Ch<>[];
                R_SUBFLAGPARITY:
                  Result:=[Ch_WParityFlag,Ch_RWParityFlag,Ch_WFlags,Ch_RWFlags]*Ch<>[];
                R_SUBFLAGAUXILIARY:
                  Result:=[Ch_WAuxiliaryFlag,Ch_RWAuxiliaryFlag,Ch_WFlags,Ch_RWFlags]*Ch<>[];
                R_SUBFLAGZERO:
                  Result:=[Ch_WZeroFlag,Ch_RWZeroFlag,Ch_WFlags,Ch_RWFlags]*Ch<>[];
                R_SUBFLAGSIGN:
                  Result:=[Ch_WSignFlag,Ch_RWSignFlag,Ch_WFlags,Ch_RWFlags]*Ch<>[];
                R_SUBFLAGOVERFLOW:
                  Result:=[Ch_WOverflowFlag,Ch_RWOverflowFlag,Ch_WFlags,Ch_RWFlags]*Ch<>[];
                R_SUBFLAGINTERRUPT:
                  Result:=[Ch_W0IntFlag,Ch_W1IntFlag,Ch_WFlags,Ch_RWFlags]*Ch<>[];
                R_SUBFLAGDIRECTION:
                  Result:=[Ch_W0DirFlag,Ch_W1DirFlag,Ch_WFlags,Ch_RWFlags]*Ch<>[];
                else
                  internalerror(2017042602);
              end;
              exit;
            end;

        case taicpu(p1).opcode of
          A_CALL:
            { We could potentially set Result to False if the register in
              question is non-volatile for the subroutine's calling convention,
              but this would require detecting the calling convention in use and
              also assuming that the routine doesn't contain malformed assembly
              language, for example... so it could only be done under -O4 as it
              would be considered a side-effect. [Kit] }
            Result := True;
          A_MOVSD:
            { special handling for SSE MOVSD }
            if (taicpu(p1).ops>0) then
              begin
                if taicpu(p1).ops<>2 then
                  internalerror(2017042703);
                Result := (taicpu(p1).oper[1]^.typ=top_reg) and RegInOp(reg,taicpu(p1).oper[1]^);
              end;
          {  VMOVSS and VMOVSD has two and three operand flavours, this cannot modelled by x86ins.dat
             so fix it here (FK)
          }
          A_VMOVSS,
          A_VMOVSD:
            begin
              Result := (taicpu(p1).ops=3) and (taicpu(p1).oper[2]^.typ=top_reg) and RegInOp(reg,taicpu(p1).oper[2]^);
              exit;
            end;
          A_IMUL:
            Result := (taicpu(p1).oper[taicpu(p1).ops-1]^.typ=top_reg) and RegInOp(reg,taicpu(p1).oper[taicpu(p1).ops-1]^);
          else
            ;
        end;
        if Result then
          exit;
        with insprop[taicpu(p1).opcode] do
          begin
            if getregtype(reg)=R_INTREGISTER then
              begin
                case getsupreg(reg) of
                  RS_EAX:
                    if [Ch_WEAX,Ch_RWEAX,Ch_MEAX]*Ch<>[] then
                      begin
                        Result := True;
                        exit
                      end;
                  RS_ECX:
                    if [Ch_WECX,Ch_RWECX,Ch_MECX]*Ch<>[] then
                      begin
                        Result := True;
                        exit
                      end;
                  RS_EDX:
                    if [Ch_WEDX,Ch_RWEDX,Ch_MEDX]*Ch<>[] then
                      begin
                        Result := True;
                        exit
                      end;
                  RS_EBX:
                    if [Ch_WEBX,Ch_RWEBX,Ch_MEBX]*Ch<>[] then
                      begin
                        Result := True;
                        exit
                      end;
                  RS_ESP:
                    if [Ch_WESP,Ch_RWESP,Ch_MESP]*Ch<>[] then
                      begin
                        Result := True;
                        exit
                      end;
                  RS_EBP:
                    if [Ch_WEBP,Ch_RWEBP,Ch_MEBP]*Ch<>[] then
                      begin
                        Result := True;
                        exit
                      end;
                  RS_ESI:
                    if [Ch_WESI,Ch_RWESI,Ch_MESI]*Ch<>[] then
                      begin
                        Result := True;
                        exit
                      end;
                  RS_EDI:
                    if [Ch_WEDI,Ch_RWEDI,Ch_MEDI]*Ch<>[] then
                      begin
                        Result := True;
                        exit
                      end;
                end;
              end;
            if ([CH_RWOP1,CH_WOP1,CH_MOP1]*Ch<>[]) and reginop(reg,taicpu(p1).oper[0]^) then
              begin
                Result := true;
                exit
              end;
            if ([Ch_RWOP2,Ch_WOP2,Ch_MOP2]*Ch<>[]) and reginop(reg,taicpu(p1).oper[1]^) then
              begin
                Result := true;
                exit
              end;
            if ([Ch_RWOP3,Ch_WOP3,Ch_MOP3]*Ch<>[]) and reginop(reg,taicpu(p1).oper[2]^) then
              begin
                Result := true;
                exit
              end;
            if ([Ch_RWOP4,Ch_WOP4,Ch_MOP4]*Ch<>[]) and reginop(reg,taicpu(p1).oper[3]^) then
              begin
                Result := true;
                exit
              end;
          end;
      end;


{$ifdef DEBUG_AOPTCPU}
    procedure TX86AsmOptimizer.DebugMsg(const s: string;p : tai);
      begin
        asml.insertbefore(tai_comment.Create(strpnew(s)), p);
      end;

    function debug_tostr(i: tcgint): string; inline;
      begin
        Result := tostr(i);
      end;

    function debug_regname(r: TRegister): string; inline;
      begin
        Result := '%' + std_regname(r);
      end;

    { Debug output function - creates a string representation of an operator }
    function debug_operstr(oper: TOper): string;
      begin
        case oper.typ of
          top_const:
            Result := '$' + debug_tostr(oper.val);
          top_reg:
            Result := debug_regname(oper.reg);
          top_ref:
            begin
              if oper.ref^.offset <> 0 then
                Result := debug_tostr(oper.ref^.offset) + '('
              else
                Result := '(';

              if (oper.ref^.base <> NR_INVALID) and (oper.ref^.base <> NR_NO) then
                begin
                  Result := Result + debug_regname(oper.ref^.base);
                  if (oper.ref^.index <> NR_INVALID) and (oper.ref^.index <> NR_NO) then
                    Result := Result + ',' + debug_regname(oper.ref^.index);
                end
              else
                if (oper.ref^.index <> NR_INVALID) and (oper.ref^.index <> NR_NO) then
                  Result := Result + debug_regname(oper.ref^.index);

              if (oper.ref^.scalefactor > 1) then
                Result := Result + ',' + debug_tostr(oper.ref^.scalefactor) + ')'
              else
                Result := Result + ')';
            end;
          else
            Result := '[UNKNOWN]';
        end;
      end;

    function debug_op2str(opcode: tasmop): string; inline;
      begin
        Result := std_op2str[opcode];
      end;

    function debug_opsize2str(opsize: topsize): string; inline;
      begin
        Result := gas_opsize2str[opsize];
      end;

{$else DEBUG_AOPTCPU}
    procedure TX86AsmOptimizer.DebugMsg(const s: string;p : tai);inline;
      begin
      end;

    function debug_tostr(i: tcgint): string; inline;
      begin
        Result := '';
      end;

    function debug_regname(r: TRegister): string; inline;
      begin
        Result := '';
      end;

    function debug_operstr(oper: TOper): string; inline;
      begin
        Result := '';
      end;

    function debug_op2str(opcode: tasmop): string; inline;
      begin
        Result := '';
      end;

    function debug_opsize2str(opsize: topsize): string; inline;
      begin
        Result := '';
      end;
{$endif DEBUG_AOPTCPU}

    class function TX86AsmOptimizer.IsMOVZXAcceptable: Boolean; inline;
      begin
{$ifdef x86_64}
        { Always fine on x86-64 }
        Result := True;
{$else x86_64}
        Result :=
{$ifdef i8086}
          (current_settings.cputype >= cpu_386) and
{$endif i8086}
          (
            { Always accept if optimising for size }
            (cs_opt_size in current_settings.optimizerswitches) or
            { From the Pentium II onwards, MOVZX only takes 1 cycle. [Kit] }
            (current_settings.optimizecputype >= cpu_Pentium2)
          );
{$endif x86_64}
      end;

    function TX86AsmOptimizer.Reg1WriteOverwritesReg2Entirely(reg1, reg2: tregister): boolean;
      begin
        if not SuperRegistersEqual(reg1,reg2) then
          exit(false);
        if getregtype(reg1)<>R_INTREGISTER then
          exit(true);  {because SuperRegisterEqual is true}
        case getsubreg(reg1) of
          { A write to R_SUBL doesn't change R_SUBH and if reg2 is R_SUBW or
            higher, it preserves the high bits, so the new value depends on
            reg2's previous value. In other words, it is equivalent to doing:

            reg2 := (reg2 and $ffffff00) or byte(reg1); }
          R_SUBL:
            exit(getsubreg(reg2)=R_SUBL);
          { A write to R_SUBH doesn't change R_SUBL and if reg2 is R_SUBW or
            higher, it actually does a:

            reg2 := (reg2 and $ffff00ff) or (reg1 and $ff00); }
          R_SUBH:
            exit(getsubreg(reg2)=R_SUBH);
          { If reg2 is R_SUBD or larger, a write to R_SUBW preserves the high 16
            bits of reg2:

            reg2 := (reg2 and $ffff0000) or word(reg1); }
          R_SUBW:
            exit(getsubreg(reg2) in [R_SUBL,R_SUBH,R_SUBW]);
          { a write to R_SUBD always overwrites every other subregister,
            because it clears the high 32 bits of R_SUBQ on x86_64 }
          R_SUBD,
          R_SUBQ:
            exit(true);
          else
            internalerror(2017042801);
        end;
      end;


    function TX86AsmOptimizer.Reg1ReadDependsOnReg2(reg1, reg2: tregister): boolean;
      begin
        if not SuperRegistersEqual(reg1,reg2) then
          exit(false);
        if getregtype(reg1)<>R_INTREGISTER then
          exit(true);  {because SuperRegisterEqual is true}
        case getsubreg(reg1) of
          R_SUBL:
            exit(getsubreg(reg2)<>R_SUBH);
          R_SUBH:
            exit(getsubreg(reg2)<>R_SUBL);
          R_SUBW,
          R_SUBD,
          R_SUBQ:
            exit(true);
          else
            internalerror(2017042802);
        end;
      end;


    function TX86AsmOptimizer.PrePeepholeOptSxx(var p : tai) : boolean;
      var
        hp1 : tai;
        l : TCGInt;
      begin
        result:=false;
        { changes the code sequence
          shr/sar const1, x
          shl     const2, x

          to

          either "sar/and", "shl/and" or just "and" depending on const1 and const2 }
        if GetNextInstruction(p, hp1) and
          MatchInstruction(hp1,A_SHL,[]) and
          (taicpu(p).oper[0]^.typ = top_const) and
          (taicpu(hp1).oper[0]^.typ = top_const) and
          (taicpu(hp1).opsize = taicpu(p).opsize) and
          (taicpu(hp1).oper[1]^.typ = taicpu(p).oper[1]^.typ) and
          OpsEqual(taicpu(hp1).oper[1]^, taicpu(p).oper[1]^) then
          begin
            if (taicpu(p).oper[0]^.val > taicpu(hp1).oper[0]^.val) and
              not(cs_opt_size in current_settings.optimizerswitches) then
              begin
                { shr/sar const1, %reg
                  shl     const2, %reg
                  with const1 > const2 }
                taicpu(p).loadConst(0,taicpu(p).oper[0]^.val-taicpu(hp1).oper[0]^.val);
                taicpu(hp1).opcode := A_AND;
                l := (1 shl (taicpu(hp1).oper[0]^.val)) - 1;
                case taicpu(p).opsize Of
                  S_B: taicpu(hp1).loadConst(0,l Xor $ff);
                  S_W: taicpu(hp1).loadConst(0,l Xor $ffff);
                  S_L: taicpu(hp1).loadConst(0,l Xor tcgint($ffffffff));
                  S_Q: taicpu(hp1).loadConst(0,l Xor tcgint($ffffffffffffffff));
                  else
                    Internalerror(2017050703)
                end;
              end
            else if (taicpu(p).oper[0]^.val<taicpu(hp1).oper[0]^.val) and
              not(cs_opt_size in current_settings.optimizerswitches) then
              begin
                { shr/sar const1, %reg
                  shl     const2, %reg
                  with const1 < const2 }
                taicpu(hp1).loadConst(0,taicpu(hp1).oper[0]^.val-taicpu(p).oper[0]^.val);
                taicpu(p).opcode := A_AND;
                l := (1 shl (taicpu(p).oper[0]^.val))-1;
                case taicpu(p).opsize Of
                  S_B: taicpu(p).loadConst(0,l Xor $ff);
                  S_W: taicpu(p).loadConst(0,l Xor $ffff);
                  S_L: taicpu(p).loadConst(0,l Xor tcgint($ffffffff));
                  S_Q: taicpu(p).loadConst(0,l Xor tcgint($ffffffffffffffff));
                  else
                    Internalerror(2017050702)
                end;
              end
            else if (taicpu(p).oper[0]^.val = taicpu(hp1).oper[0]^.val) then
              begin
                { shr/sar const1, %reg
                  shl     const2, %reg
                  with const1 = const2 }
                taicpu(p).opcode := A_AND;
                l := (1 shl (taicpu(p).oper[0]^.val))-1;
                case taicpu(p).opsize Of
                  S_B: taicpu(p).loadConst(0,l Xor $ff);
                  S_W: taicpu(p).loadConst(0,l Xor $ffff);
                  S_L: taicpu(p).loadConst(0,l Xor tcgint($ffffffff));
                  S_Q: taicpu(p).loadConst(0,l Xor tcgint($ffffffffffffffff));
                  else
                    Internalerror(2017050701)
                end;
                RemoveInstruction(hp1);
              end;
          end;
      end;


    function TX86AsmOptimizer.PrePeepholeOptIMUL(var p : tai) : boolean;
      var
        opsize : topsize;
        hp1 : tai;
        tmpref : treference;
        ShiftValue : Cardinal;
        BaseValue : TCGInt;
      begin
        result:=false;
        opsize:=taicpu(p).opsize;
        { changes certain "imul const, %reg"'s to lea sequences }
        if (MatchOpType(taicpu(p),top_const,top_reg) or
            MatchOpType(taicpu(p),top_const,top_reg,top_reg)) and
           (opsize in [S_L{$ifdef x86_64},S_Q{$endif x86_64}]) then
          if (taicpu(p).oper[0]^.val = 1) then
            if (taicpu(p).ops = 2) then
             { remove "imul $1, reg" }
              begin
                DebugMsg(SPeepholeOptimization + 'Imul2Nop done',p);
                Result := RemoveCurrentP(p);
              end
            else
             { change "imul $1, reg1, reg2" to "mov reg1, reg2" }
              begin
                hp1 := taicpu.Op_Reg_Reg(A_MOV, opsize, taicpu(p).oper[1]^.reg,taicpu(p).oper[2]^.reg);
                InsertLLItem(p.previous, p.next, hp1);
                DebugMsg(SPeepholeOptimization + 'Imul2Mov done',p);
                p.free;
                p := hp1;
              end
          else if ((taicpu(p).ops <= 2) or
              (taicpu(p).oper[2]^.typ = Top_Reg)) and
           not(cs_opt_size in current_settings.optimizerswitches) and
           (not(GetNextInstruction(p, hp1)) or
             not((tai(hp1).typ = ait_instruction) and
                 ((taicpu(hp1).opcode=A_Jcc) and
                  (taicpu(hp1).condition in [C_O,C_NO])))) then
            begin
              {
                imul X, reg1, reg2 to
                  lea (reg1,reg1,Y), reg2
                  shl ZZ,reg2
                imul XX, reg1 to
                  lea (reg1,reg1,YY), reg1
                  shl ZZ,reg2

                This optimziation makes sense for pretty much every x86, except the VIA Nano3000: it has IMUL latency 2, lea/shl pair as well,
                it does not exist as a separate optimization target in FPC though.

                This optimziation can be applied as long as only two bits are set in the constant and those two bits are separated by
                at most two zeros
              }
              reference_reset(tmpref,1,[]);
              if (PopCnt(QWord(taicpu(p).oper[0]^.val))=2) and (BsrQWord(taicpu(p).oper[0]^.val)-BsfQWord(taicpu(p).oper[0]^.val)<=3) then
                begin
                  ShiftValue:=BsfQWord(taicpu(p).oper[0]^.val);
                  BaseValue:=taicpu(p).oper[0]^.val shr ShiftValue;
                  TmpRef.base := taicpu(p).oper[1]^.reg;
                  TmpRef.index := taicpu(p).oper[1]^.reg;
                  if not(BaseValue in [3,5,9]) then
                    Internalerror(2018110101);
                  TmpRef.ScaleFactor := BaseValue-1;
                  if (taicpu(p).ops = 2) then
                    hp1 := taicpu.op_ref_reg(A_LEA, opsize, TmpRef, taicpu(p).oper[1]^.reg)
                  else
                    hp1 := taicpu.op_ref_reg(A_LEA, opsize, TmpRef, taicpu(p).oper[2]^.reg);
                  AsmL.InsertAfter(hp1,p);
                  DebugMsg(SPeepholeOptimization + 'Imul2LeaShl done',p);
                  taicpu(hp1).fileinfo:=taicpu(p).fileinfo;
                  RemoveCurrentP(p, hp1);
                  if ShiftValue>0 then
                    AsmL.InsertAfter(taicpu.op_const_reg(A_SHL, opsize, ShiftValue, taicpu(hp1).oper[1]^.reg),hp1);
              end;
            end;
      end;


    function TX86AsmOptimizer.RegLoadedWithNewValue(reg: tregister; hp: tai): boolean;
      var
        p: taicpu;
      begin
        if not assigned(hp) or
           (hp.typ <> ait_instruction) then
         begin
           Result := false;
           exit;
         end;
        p := taicpu(hp);
        if SuperRegistersEqual(reg,NR_DEFAULTFLAGS) then
          with insprop[p.opcode] do
            begin
              case getsubreg(reg) of
                R_SUBW,R_SUBD,R_SUBQ:
                  Result:=
                    RegLoadedWithNewValue(NR_CARRYFLAG,hp) and
                    RegLoadedWithNewValue(NR_PARITYFLAG,hp) and
                    RegLoadedWithNewValue(NR_AUXILIARYFLAG,hp) and
                    RegLoadedWithNewValue(NR_ZEROFLAG,hp) and
                    RegLoadedWithNewValue(NR_SIGNFLAG,hp) and
                    RegLoadedWithNewValue(NR_OVERFLOWFLAG,hp);
                R_SUBFLAGCARRY:
                  Result:=[Ch_W0CarryFlag,Ch_W1CarryFlag,Ch_WCarryFlag,Ch_WUCarryFlag,Ch_WFlags]*Ch<>[];
                R_SUBFLAGPARITY:
                  Result:=[Ch_W0ParityFlag,Ch_W1ParityFlag,Ch_WParityFlag,Ch_WUParityFlag,Ch_WFlags]*Ch<>[];
                R_SUBFLAGAUXILIARY:
                  Result:=[Ch_W0AuxiliaryFlag,Ch_W1AuxiliaryFlag,Ch_WAuxiliaryFlag,Ch_WUAuxiliaryFlag,Ch_WFlags]*Ch<>[];
                R_SUBFLAGZERO:
                  Result:=[Ch_W0ZeroFlag,Ch_W1ZeroFlag,Ch_WZeroFlag,Ch_WUZeroFlag,Ch_WFlags]*Ch<>[];
                R_SUBFLAGSIGN:
                  Result:=[Ch_W0SignFlag,Ch_W1SignFlag,Ch_WSignFlag,Ch_WUSignFlag,Ch_WFlags]*Ch<>[];
                R_SUBFLAGOVERFLOW:
                  Result:=[Ch_W0OverflowFlag,Ch_W1OverflowFlag,Ch_WOverflowFlag,Ch_WUOverflowFlag,Ch_WFlags]*Ch<>[];
                R_SUBFLAGINTERRUPT:
                  Result:=[Ch_W0IntFlag,Ch_W1IntFlag,Ch_WFlags]*Ch<>[];
                R_SUBFLAGDIRECTION:
                  Result:=[Ch_W0DirFlag,Ch_W1DirFlag,Ch_WFlags]*Ch<>[];
                else
                  begin
                  writeln(getsubreg(reg));
                  internalerror(2017050501);
                  end;
              end;
              exit;
            end;
        Result :=
          (((p.opcode = A_MOV) or
            (p.opcode = A_MOVZX) or
            (p.opcode = A_MOVSX) or
            (p.opcode = A_LEA) or
            (p.opcode = A_VMOVSS) or
            (p.opcode = A_VMOVSD) or
            (p.opcode = A_VMOVAPD) or
            (p.opcode = A_VMOVAPS) or
            (p.opcode = A_VMOVQ) or
            (p.opcode = A_MOVSS) or
            (p.opcode = A_MOVSD) or
            (p.opcode = A_MOVQ) or
            (p.opcode = A_MOVAPD) or
            (p.opcode = A_MOVAPS) or
{$ifndef x86_64}
            (p.opcode = A_LDS) or
            (p.opcode = A_LES) or
{$endif not x86_64}
            (p.opcode = A_LFS) or
            (p.opcode = A_LGS) or
            (p.opcode = A_LSS)) and
           (p.ops=2) and  { A_MOVSD can have zero operands, so this check is needed }
           (p.oper[1]^.typ = top_reg) and
           (Reg1WriteOverwritesReg2Entirely(p.oper[1]^.reg,reg)) and
           ((p.oper[0]^.typ = top_const) or
            ((p.oper[0]^.typ = top_reg) and
             not(Reg1ReadDependsOnReg2(p.oper[0]^.reg,reg))) or
            ((p.oper[0]^.typ = top_ref) and
             not RegInRef(reg,p.oper[0]^.ref^)))) or
          ((p.opcode = A_POP) and
           (Reg1WriteOverwritesReg2Entirely(p.oper[0]^.reg,reg))) or
          ((p.opcode = A_IMUL) and
           (p.ops=3) and
           (Reg1WriteOverwritesReg2Entirely(p.oper[2]^.reg,reg)) and
           (((p.oper[1]^.typ=top_reg) and not(Reg1ReadDependsOnReg2(p.oper[1]^.reg,reg))) or
            ((p.oper[1]^.typ=top_ref) and not(RegInRef(reg,p.oper[1]^.ref^))))) or
          ((((p.opcode = A_IMUL) or
             (p.opcode = A_MUL)) and
            (p.ops=1)) and
           (((p.oper[0]^.typ=top_reg) and not(Reg1ReadDependsOnReg2(p.oper[0]^.reg,reg))) or
            ((p.oper[0]^.typ=top_ref) and not(RegInRef(reg,p.oper[0]^.ref^)))) and
           (((p.opsize=S_B) and Reg1WriteOverwritesReg2Entirely(NR_AX,reg) and not(Reg1ReadDependsOnReg2(NR_AL,reg))) or
            ((p.opsize=S_W) and Reg1WriteOverwritesReg2Entirely(NR_DX,reg)) or
            ((p.opsize=S_L) and Reg1WriteOverwritesReg2Entirely(NR_EDX,reg))
{$ifdef x86_64}
         or ((p.opsize=S_Q) and Reg1WriteOverwritesReg2Entirely(NR_RDX,reg))
{$endif x86_64}
           )) or
          ((p.opcode = A_CWD) and Reg1WriteOverwritesReg2Entirely(NR_DX,reg)) or
          ((p.opcode = A_CDQ) and Reg1WriteOverwritesReg2Entirely(NR_EDX,reg)) or
{$ifdef x86_64}
          ((p.opcode = A_CQO) and Reg1WriteOverwritesReg2Entirely(NR_RDX,reg)) or
{$endif x86_64}
          ((p.opcode = A_CBW) and Reg1WriteOverwritesReg2Entirely(NR_AX,reg) and not(Reg1ReadDependsOnReg2(NR_AL,reg))) or
{$ifndef x86_64}
          ((p.opcode = A_LDS) and (reg=NR_DS) and not(RegInRef(reg,p.oper[0]^.ref^))) or
          ((p.opcode = A_LES) and (reg=NR_ES) and not(RegInRef(reg,p.oper[0]^.ref^))) or
{$endif not x86_64}
          ((p.opcode = A_LFS) and (reg=NR_FS) and not(RegInRef(reg,p.oper[0]^.ref^))) or
          ((p.opcode = A_LGS) and (reg=NR_GS) and not(RegInRef(reg,p.oper[0]^.ref^))) or
          ((p.opcode = A_LSS) and (reg=NR_SS) and not(RegInRef(reg,p.oper[0]^.ref^))) or
{$ifndef x86_64}
          ((p.opcode = A_AAM) and Reg1WriteOverwritesReg2Entirely(NR_AH,reg)) or
{$endif not x86_64}
          ((p.opcode = A_LAHF) and Reg1WriteOverwritesReg2Entirely(NR_AH,reg)) or
          ((p.opcode = A_LODSB) and Reg1WriteOverwritesReg2Entirely(NR_AL,reg)) or
          ((p.opcode = A_LODSW) and Reg1WriteOverwritesReg2Entirely(NR_AX,reg)) or
          ((p.opcode = A_LODSD) and Reg1WriteOverwritesReg2Entirely(NR_EAX,reg)) or
{$ifdef x86_64}
          ((p.opcode = A_LODSQ) and Reg1WriteOverwritesReg2Entirely(NR_RAX,reg)) or
{$endif x86_64}
          ((p.opcode = A_SETcc) and (p.oper[0]^.typ=top_reg) and Reg1WriteOverwritesReg2Entirely(p.oper[0]^.reg,reg)) or
          (((p.opcode = A_FSTSW) or
            (p.opcode = A_FNSTSW)) and
           (p.oper[0]^.typ=top_reg) and
           Reg1WriteOverwritesReg2Entirely(p.oper[0]^.reg,reg)) or
          (((p.opcode = A_XOR) or (p.opcode = A_SUB) or (p.opcode = A_SBB)) and
           (p.oper[0]^.typ=top_reg) and (p.oper[1]^.typ=top_reg) and
           (p.oper[0]^.reg=p.oper[1]^.reg) and
           Reg1WriteOverwritesReg2Entirely(p.oper[1]^.reg,reg));
      end;


    class function TX86AsmOptimizer.IsExitCode(p : tai) : boolean;
      var
        hp2,hp3 : tai;
      begin
        { some x86-64 issue a NOP before the real exit code }
        if MatchInstruction(p,A_NOP,[]) then
          GetNextInstruction(p,p);
        result:=assigned(p) and (p.typ=ait_instruction) and
        ((taicpu(p).opcode = A_RET) or
         ((taicpu(p).opcode=A_LEAVE) and
          GetNextInstruction(p,hp2) and
          MatchInstruction(hp2,A_RET,[S_NO])
         ) or
         (((taicpu(p).opcode=A_LEA) and
           MatchOpType(taicpu(p),top_ref,top_reg) and
           (taicpu(p).oper[0]^.ref^.base=NR_STACK_POINTER_REG) and
           (taicpu(p).oper[1]^.reg=NR_STACK_POINTER_REG)
           ) and
          GetNextInstruction(p,hp2) and
          MatchInstruction(hp2,A_RET,[S_NO])
         ) or
         ((((taicpu(p).opcode=A_MOV) and
           MatchOpType(taicpu(p),top_reg,top_reg) and
           (taicpu(p).oper[0]^.reg=current_procinfo.framepointer) and
           (taicpu(p).oper[1]^.reg=NR_STACK_POINTER_REG)) or
           ((taicpu(p).opcode=A_LEA) and
           MatchOpType(taicpu(p),top_ref,top_reg) and
           (taicpu(p).oper[0]^.ref^.base=current_procinfo.framepointer) and
           (taicpu(p).oper[1]^.reg=NR_STACK_POINTER_REG)
           )
          ) and
          GetNextInstruction(p,hp2) and
          MatchInstruction(hp2,A_POP,[reg2opsize(current_procinfo.framepointer)]) and
          MatchOpType(taicpu(hp2),top_reg) and
          (taicpu(hp2).oper[0]^.reg=current_procinfo.framepointer) and
          GetNextInstruction(hp2,hp3) and
          MatchInstruction(hp3,A_RET,[S_NO])
         )
        );
      end;


    class function TX86AsmOptimizer.isFoldableArithOp(hp1: taicpu; reg: tregister): boolean;
      begin
        isFoldableArithOp := False;
        case hp1.opcode of
          A_ADD,A_SUB,A_OR,A_XOR,A_AND,A_SHL,A_SHR,A_SAR:
            isFoldableArithOp :=
              ((taicpu(hp1).oper[0]^.typ = top_const) or
               ((taicpu(hp1).oper[0]^.typ = top_reg) and
                (taicpu(hp1).oper[0]^.reg <> reg))) and
              (taicpu(hp1).oper[1]^.typ = top_reg) and
              (taicpu(hp1).oper[1]^.reg = reg);
          A_INC,A_DEC,A_NEG,A_NOT:
            isFoldableArithOp :=
              (taicpu(hp1).oper[0]^.typ = top_reg) and
              (taicpu(hp1).oper[0]^.reg = reg);
          else
            ;
        end;
      end;


    procedure TX86AsmOptimizer.RemoveLastDeallocForFuncRes(p: tai);

      procedure DoRemoveLastDeallocForFuncRes( supreg: tsuperregister);
        var
          hp2: tai;
        begin
          hp2 := p;
          repeat
            hp2 := tai(hp2.previous);
            if assigned(hp2) and
               (hp2.typ = ait_regalloc) and
               (tai_regalloc(hp2).ratype=ra_dealloc) and
               (getregtype(tai_regalloc(hp2).reg) = R_INTREGISTER) and
               (getsupreg(tai_regalloc(hp2).reg) = supreg) then
              begin
                RemoveInstruction(hp2);
                break;
              end;
          until not(assigned(hp2)) or regInInstruction(newreg(R_INTREGISTER,supreg,R_SUBWHOLE),hp2);
        end;

      begin
          case current_procinfo.procdef.returndef.typ of
            arraydef,recorddef,pointerdef,
               stringdef,enumdef,procdef,objectdef,errordef,
               filedef,setdef,procvardef,
               classrefdef,forwarddef:
              DoRemoveLastDeallocForFuncRes(RS_EAX);
            orddef:
              if current_procinfo.procdef.returndef.size <> 0 then
                begin
                  DoRemoveLastDeallocForFuncRes(RS_EAX);
                  { for int64/qword }
                  if current_procinfo.procdef.returndef.size = 8 then
                    DoRemoveLastDeallocForFuncRes(RS_EDX);
                end;
            else
              ;
          end;
      end;


    function TX86AsmOptimizer.OptPass1_V_MOVAP(var p : tai) : boolean;
      var
        hp1,hp2 : tai;
      begin
        result:=false;
        if MatchOpType(taicpu(p),top_reg,top_reg) then
          begin
            { vmova* reg1,reg1
              =>
              <nop> }
            if MatchOperand(taicpu(p).oper[0]^,taicpu(p).oper[1]^) then
              begin
                RemoveCurrentP(p);
                result:=true;
                exit;
              end
            else if GetNextInstruction(p,hp1) then
              begin
                if MatchInstruction(hp1,[taicpu(p).opcode],[S_NO]) and
                  MatchOpType(taicpu(hp1),top_reg,top_reg) and
                  MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[0]^) then
                  begin
                    { vmova* reg1,reg2
                      vmova* reg2,reg3
                      dealloc reg2
                      =>
                      vmova* reg1,reg3 }
                    TransferUsedRegs(TmpUsedRegs);
                    UpdateUsedRegs(TmpUsedRegs, tai(p.next));
                    if not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp1,TmpUsedRegs)) then
                      begin
                        DebugMsg(SPeepholeOptimization + '(V)MOVA*(V)MOVA*2(V)MOVA* 1',p);
                        taicpu(p).loadoper(1,taicpu(hp1).oper[1]^);
                        RemoveInstruction(hp1);
                        result:=true;
                        exit;
                      end
                    { special case:
                      vmova* reg1,reg2
                      vmova* reg2,reg1
                      =>
                      vmova* reg1,reg2 }
                    else if MatchOperand(taicpu(p).oper[0]^,taicpu(hp1).oper[1]^) then
                      begin
                        DebugMsg(SPeepholeOptimization + '(V)MOVA*(V)MOVA*2(V)MOVA* 2',p);
                        RemoveInstruction(hp1);
                        result:=true;
                        exit;
                      end
                  end
              else if ((MatchInstruction(p,[A_MOVAPS,A_VMOVAPS],[S_NO]) and
                 MatchInstruction(hp1,[A_MOVSS,A_VMOVSS],[S_NO])) or
                 ((MatchInstruction(p,[A_MOVAPD,A_VMOVAPD],[S_NO]) and
                 MatchInstruction(hp1,[A_MOVSD,A_VMOVSD],[S_NO])))
                ) and
                MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[0]^) then
                begin
                  { vmova* reg1,reg2
                    vmovs* reg2,<op>
                    dealloc reg2
                    =>
                    vmovs* reg1,reg3 }
                  TransferUsedRegs(TmpUsedRegs);
                  UpdateUsedRegs(TmpUsedRegs, tai(p.next));
                  if not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp1,TmpUsedRegs)) then
                    begin
                      DebugMsg(SPeepholeOptimization + '(V)MOVA*(V)MOVS*2(V)MOVS* 1',p);
                      taicpu(p).opcode:=taicpu(hp1).opcode;
                      taicpu(p).loadoper(1,taicpu(hp1).oper[1]^);
                      RemoveInstruction(hp1);
                      result:=true;
                      exit;
                    end
                end;
            end;
          if GetNextInstructionUsingReg(p,hp1,taicpu(p).oper[1]^.reg) then
            begin
              if MatchInstruction(hp1,[A_VFMADDPD,
                                              A_VFMADD132PD,
                                              A_VFMADD132PS,
                                              A_VFMADD132SD,
                                              A_VFMADD132SS,
                                              A_VFMADD213PD,
                                              A_VFMADD213PS,
                                              A_VFMADD213SD,
                                              A_VFMADD213SS,
                                              A_VFMADD231PD,
                                              A_VFMADD231PS,
                                              A_VFMADD231SD,
                                              A_VFMADD231SS,
                                              A_VFMADDSUB132PD,
                                              A_VFMADDSUB132PS,
                                              A_VFMADDSUB213PD,
                                              A_VFMADDSUB213PS,
                                              A_VFMADDSUB231PD,
                                              A_VFMADDSUB231PS,
                                              A_VFMSUB132PD,
                                              A_VFMSUB132PS,
                                              A_VFMSUB132SD,
                                              A_VFMSUB132SS,
                                              A_VFMSUB213PD,
                                              A_VFMSUB213PS,
                                              A_VFMSUB213SD,
                                              A_VFMSUB213SS,
                                              A_VFMSUB231PD,
                                              A_VFMSUB231PS,
                                              A_VFMSUB231SD,
                                              A_VFMSUB231SS,
                                              A_VFMSUBADD132PD,
                                              A_VFMSUBADD132PS,
                                              A_VFMSUBADD213PD,
                                              A_VFMSUBADD213PS,
                                              A_VFMSUBADD231PD,
                                              A_VFMSUBADD231PS,
                                              A_VFNMADD132PD,
                                              A_VFNMADD132PS,
                                              A_VFNMADD132SD,
                                              A_VFNMADD132SS,
                                              A_VFNMADD213PD,
                                              A_VFNMADD213PS,
                                              A_VFNMADD213SD,
                                              A_VFNMADD213SS,
                                              A_VFNMADD231PD,
                                              A_VFNMADD231PS,
                                              A_VFNMADD231SD,
                                              A_VFNMADD231SS,
                                              A_VFNMSUB132PD,
                                              A_VFNMSUB132PS,
                                              A_VFNMSUB132SD,
                                              A_VFNMSUB132SS,
                                              A_VFNMSUB213PD,
                                              A_VFNMSUB213PS,
                                              A_VFNMSUB213SD,
                                              A_VFNMSUB213SS,
                                              A_VFNMSUB231PD,
                                              A_VFNMSUB231PS,
                                              A_VFNMSUB231SD,
                                              A_VFNMSUB231SS],[S_NO]) and
                  { we mix single and double opperations here because we assume that the compiler
                    generates vmovapd only after double operations and vmovaps only after single operations }
                  MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[2]^) and
                  GetNextInstruction(hp1,hp2) and
                  MatchInstruction(hp2,[A_VMOVAPD,A_VMOVAPS,A_MOVAPD,A_MOVAPS],[S_NO]) and
                  MatchOperand(taicpu(p).oper[0]^,taicpu(hp2).oper[1]^) then
                  begin
                    TransferUsedRegs(TmpUsedRegs);
                    UpdateUsedRegs(TmpUsedRegs, tai(p.next));
                    UpdateUsedRegs(TmpUsedRegs, tai(hp1.next));
                    if not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp2,TmpUsedRegs)) then
                      begin
                        taicpu(hp1).loadoper(2,taicpu(p).oper[0]^);
                        RemoveCurrentP(p, hp1); // <-- Is this actually safe? hp1 is not necessarily the next instruction. [Kit]
                        RemoveInstruction(hp2);
                      end;
                  end
                else if (hp1.typ = ait_instruction) and
                  GetNextInstruction(hp1, hp2) and
                  MatchInstruction(hp2,taicpu(p).opcode,[]) and
                  OpsEqual(taicpu(hp2).oper[1]^, taicpu(p).oper[0]^) and
                  MatchOpType(taicpu(hp2),top_reg,top_reg) and
                  MatchOperand(taicpu(hp2).oper[0]^,taicpu(p).oper[1]^) and
                  (((taicpu(p).opcode=A_MOVAPS) and
                    ((taicpu(hp1).opcode=A_ADDSS) or (taicpu(hp1).opcode=A_SUBSS) or
                     (taicpu(hp1).opcode=A_MULSS) or (taicpu(hp1).opcode=A_DIVSS))) or
                   ((taicpu(p).opcode=A_MOVAPD) and
                    ((taicpu(hp1).opcode=A_ADDSD) or (taicpu(hp1).opcode=A_SUBSD) or
                     (taicpu(hp1).opcode=A_MULSD) or (taicpu(hp1).opcode=A_DIVSD)))
                  ) then
                  { change
                             movapX    reg,reg2
                             addsX/subsX/... reg3, reg2
                             movapX    reg2,reg
                    to
                             addsX/subsX/... reg3,reg
                  }
                  begin
                    TransferUsedRegs(TmpUsedRegs);
                    UpdateUsedRegs(TmpUsedRegs, tai(p.next));
                    UpdateUsedRegs(TmpUsedRegs, tai(hp1.next));
                    If not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp2,TmpUsedRegs)) then
                      begin
                        DebugMsg(SPeepholeOptimization + 'MovapXOpMovapX2Op ('+
                              debug_op2str(taicpu(p).opcode)+' '+
                              debug_op2str(taicpu(hp1).opcode)+' '+
                              debug_op2str(taicpu(hp2).opcode)+') done',p);
                        { we cannot eliminate the first move if
                          the operations uses the same register for source and dest }
                        if not(OpsEqual(taicpu(hp1).oper[1]^,taicpu(hp1).oper[0]^)) then
                          RemoveCurrentP(p, nil);
                        p:=hp1;
                        taicpu(hp1).loadoper(1, taicpu(hp2).oper[1]^);
                        RemoveInstruction(hp2);
                        result:=true;
                      end;
                  end;
              end;
          end;
      end;


    function TX86AsmOptimizer.OptPass1VOP(var p : tai) : boolean;
      var
        hp1 : tai;
      begin
        result:=false;
        { replace
            V<Op>X   %mreg1,%mreg2,%mreg3
            VMovX    %mreg3,%mreg4
            dealloc  %mreg3

            by
            V<Op>X   %mreg1,%mreg2,%mreg4
          ?
        }
        if GetNextInstruction(p,hp1) and
          { we mix single and double operations here because we assume that the compiler
            generates vmovapd only after double operations and vmovaps only after single operations }
          MatchInstruction(hp1,A_VMOVAPD,A_VMOVAPS,[S_NO]) and
          MatchOperand(taicpu(p).oper[2]^,taicpu(hp1).oper[0]^) and
          (taicpu(hp1).oper[1]^.typ=top_reg) then
          begin
            TransferUsedRegs(TmpUsedRegs);
            UpdateUsedRegs(TmpUsedRegs, tai(p.next));
            if not(RegUsedAfterInstruction(taicpu(hp1).oper[0]^.reg,hp1,TmpUsedRegs)) then
              begin
                taicpu(p).loadoper(2,taicpu(hp1).oper[1]^);
                DebugMsg(SPeepholeOptimization + 'VOpVmov2VOp done',p);
                RemoveInstruction(hp1);
                result:=true;
              end;
          end;
      end;


    { Replaces all references to AOldReg in a memory reference to ANewReg }
    class function TX86AsmOptimizer.ReplaceRegisterInRef(var ref: TReference; const AOldReg, ANewReg: TRegister): Boolean;
      var
        OldSupReg: TSuperRegister;
        OldSubReg, MemSubReg: TSubRegister;
      begin
        Result := False;
        { For safety reasons, only check for exact register matches }

        { Check base register }
        if (ref.base = AOldReg) then
          begin
            ref.base := ANewReg;
            Result := True;
          end;

        { Check index register }
        if (ref.index = AOldReg) then
          begin
            ref.index := ANewReg;
            Result := True;
          end;
      end;


    { Replaces all references to AOldReg in an operand to ANewReg }
    class function TX86AsmOptimizer.ReplaceRegisterInOper(const p: taicpu; const OperIdx: Integer; const AOldReg, ANewReg: TRegister): Boolean;
      var
        OldSupReg, NewSupReg: TSuperRegister;
        OldSubReg, NewSubReg, MemSubReg: TSubRegister;
        OldRegType: TRegisterType;
        ThisOper: POper;
      begin
        ThisOper := p.oper[OperIdx]; { Faster to access overall }
        Result := False;

        if (AOldReg = NR_NO) or (ANewReg = NR_NO) then
          InternalError(2020011801);

        OldSupReg := getsupreg(AOldReg);
        OldSubReg := getsubreg(AOldReg);
        OldRegType := getregtype(AOldReg);
        NewSupReg := getsupreg(ANewReg);
        NewSubReg := getsubreg(ANewReg);

        if OldRegType <> getregtype(ANewReg) then
          InternalError(2020011802);

        if OldSubReg <> NewSubReg then
          InternalError(2020011803);

        case ThisOper^.typ of
          top_reg:
            if (
              (ThisOper^.reg = AOldReg) or
                (
                  (OldRegType = R_INTREGISTER) and
                  (getsupreg(ThisOper^.reg) = OldSupReg) and
                  (getregtype(ThisOper^.reg) = R_INTREGISTER) and
                  (
                    (getsubreg(ThisOper^.reg) <= OldSubReg)
{$ifndef x86_64}
                    and (
                    { Under i386 and i8086, ESI, EDI, EBP and ESP
                      don't have an 8-bit representation }
                      (getsubreg(ThisOper^.reg) >= R_SUBW) or
                      not (NewSupReg in [RS_ESI, RS_EDI, RS_EBP, RS_ESP])
                    )
{$endif x86_64}
                  )
                )
              ) then
              begin
                ThisOper^.reg := newreg(getregtype(ANewReg), NewSupReg, getsubreg(p.oper[OperIdx]^.reg));
                Result := True;
              end;
          top_ref:
            if ReplaceRegisterInRef(ThisOper^.ref^, AOldReg, ANewReg) then
              Result := True;

          else
            ;
        end;
      end;


    { Replaces all references to AOldReg in an instruction to ANewReg }
    function TX86AsmOptimizer.ReplaceRegisterInInstruction(const p: taicpu; const AOldReg, ANewReg: TRegister): Boolean;
      const
        ReadFlag: array[0..3] of TInsChange = (Ch_Rop1, Ch_Rop2, Ch_Rop3, Ch_Rop4);
      var
        OperIdx: Integer;
      begin
        Result := False;

        for OperIdx := 0 to p.ops - 1 do
          if (ReadFlag[OperIdx] in InsProp[p.Opcode].Ch) and
          { The shift and rotate instructions can only use CL }
          not (
            (OperIdx = 0) and
            { This second condition just helps to avoid unnecessarily
              calling MatchInstruction for 10 different opcodes }
            (p.oper[0]^.reg = NR_CL) and
            MatchInstruction(p, [A_RCL, A_RCR, A_ROL, A_ROR, A_SAL, A_SAR, A_SHL, A_SHLD, A_SHR, A_SHRD], [])
          ) then
            Result := ReplaceRegisterInOper(p, OperIdx, AOldReg, ANewReg) or Result;
      end;


    class function TX86AsmOptimizer.IsRefSafe(const ref: PReference): Boolean; inline;
      begin
        Result :=
          (ref^.index = NR_NO) and
          (
{$ifdef x86_64}
            (
              (ref^.base = NR_RIP) and
              (ref^.refaddr in [addr_pic, addr_pic_no_got])
            ) or
{$endif x86_64}
            (ref^.base = NR_STACK_POINTER_REG) or
            (ref^.base = current_procinfo.framepointer)
          );
      end;


    function TX86AsmOptimizer.ConvertLEA(const p: taicpu): Boolean;
      var
        l: asizeint;
      begin
        Result := False;

        { Should have been checked previously }
        if p.opcode <> A_LEA then
          InternalError(2020072501);

        { do not mess with the stack point as adjusting it by lea is recommend, except if we optimize for size }
         if (p.oper[1]^.reg=NR_STACK_POINTER_REG) and
           not(cs_opt_size in current_settings.optimizerswitches) then
           exit;

         with p.oper[0]^.ref^ do
          begin
            if (base <> p.oper[1]^.reg) or
               (index <> NR_NO) or
               assigned(symbol) then
              exit;

            l:=offset;
            if (l=1) and UseIncDec then
              begin
                p.opcode:=A_INC;
                p.loadreg(0,p.oper[1]^.reg);
                p.ops:=1;
                DebugMsg(SPeepholeOptimization + 'Lea2Inc done',p);
              end
            else if (l=-1) and UseIncDec then
              begin
                p.opcode:=A_DEC;
                p.loadreg(0,p.oper[1]^.reg);
                p.ops:=1;
                DebugMsg(SPeepholeOptimization + 'Lea2Dec done',p);
              end
            else
              begin
                if (l<0) and (l<>-2147483648) then
                  begin
                    p.opcode:=A_SUB;
                    p.loadConst(0,-l);
                    DebugMsg(SPeepholeOptimization + 'Lea2Sub done',p);
                  end
                else
                  begin
                    p.opcode:=A_ADD;
                    p.loadConst(0,l);
                    DebugMsg(SPeepholeOptimization + 'Lea2Add done',p);
                  end;
              end;
          end;

        Result := True;
      end;


    function TX86AsmOptimizer.DeepMOVOpt(const p_mov: taicpu; const hp: taicpu): Boolean;
      var
        CurrentReg, ReplaceReg: TRegister;
        SubReg: TSubRegister;
      begin
        Result := False;

        ReplaceReg := taicpu(p_mov).oper[0]^.reg;
        CurrentReg := taicpu(p_mov).oper[1]^.reg;

        case hp.opcode of
          A_FSTSW, A_FNSTSW,
          A_IN,   A_INS,  A_OUT,  A_OUTS,
          A_CMPS, A_LODS, A_MOVS, A_SCAS, A_STOS:
            { These routines have explicit operands, but they are restricted in
              what they can be (e.g. IN and OUT can only read from AL, AX or
              EAX. }
            Exit;

          A_IMUL:
            begin
                { The 1-operand version writes to implicit registers
                  The 2-operand version reads from the first operator, and reads
                  from and writes to the second (equivalent to Ch_ROp1, ChRWOp2).
                  the 3-operand version reads from a register that it doesn't write to
                }
              case hp.ops of
                1:
                  if (
                    (
                      (hp.opsize = S_B) and (getsupreg(CurrentReg) <> RS_EAX)
                    ) or
                      not (getsupreg(CurrentReg) in [RS_EAX, RS_EDX])
                  ) and ReplaceRegisterInOper(hp, 0, CurrentReg, ReplaceReg) then
                    begin
                      Result := True;
                      DebugMsg(SPeepholeOptimization + debug_regname(CurrentReg) + ' = ' + debug_regname(ReplaceReg) + '; changed to minimise pipeline stall (MovIMul2MovIMul 1)', hp);
                      AllocRegBetween(ReplaceReg, p_mov, hp, UsedRegs);
                    end;
                2:
                  { Only modify the first parameter }
                  if ReplaceRegisterInOper(hp, 0, CurrentReg, ReplaceReg) then
                    begin
                      Result := True;
                      DebugMsg(SPeepholeOptimization + debug_regname(CurrentReg) + ' = ' + debug_regname(ReplaceReg) + '; changed to minimise pipeline stall (MovIMul2MovIMul 2)', hp);
                      AllocRegBetween(ReplaceReg, p_mov, hp, UsedRegs);
                    end;
                3:
                  { Only modify the second parameter }
                  if ReplaceRegisterInOper(hp, 1, CurrentReg, ReplaceReg) then
                    begin
                      Result := True;
                      DebugMsg(SPeepholeOptimization + debug_regname(CurrentReg) + ' = ' + debug_regname(ReplaceReg) + '; changed to minimise pipeline stall (MovIMul2MovIMul 3)', hp);
                      AllocRegBetween(ReplaceReg, p_mov, hp, UsedRegs);
                    end;
                else
                  InternalError(2020012901);
              end;
            end;

          else
            if (hp.ops > 0) and
              ReplaceRegisterInInstruction(hp, CurrentReg, ReplaceReg) then
              begin
                Result := True;
                DebugMsg(SPeepholeOptimization + debug_regname(CurrentReg) + ' = ' + debug_regname(ReplaceReg) + '; changed to minimise pipeline stall (MovXXX2MovXXX)', hp);

                AllocRegBetween(ReplaceReg, p_mov, hp, UsedRegs);
              end;
          end;
      end;


    function TX86AsmOptimizer.OptPass1MOV(var p : tai) : boolean;
    var
      hp1, hp2, hp3: tai;

      procedure convert_mov_value(signed_movop: tasmop; max_value: tcgint); inline;
        begin
          if taicpu(hp1).opcode = signed_movop then
            begin
              if taicpu(p).oper[0]^.val > max_value shr 1 then
                taicpu(p).oper[0]^.val:=taicpu(p).oper[0]^.val - max_value - 1 { Convert to signed }
            end
          else
            taicpu(p).oper[0]^.val:=taicpu(p).oper[0]^.val and max_value; { Trim to unsigned }
        end;

      var
        GetNextInstruction_p, TempRegUsed: Boolean;
        PreMessage, RegName1, RegName2, InputVal, MaskNum: string;
        NewSize: topsize;
        CurrentReg: TRegister;
      begin
        Result:=false;

        GetNextInstruction_p:=GetNextInstruction(p, hp1);

        {  remove mov reg1,reg1? }
        if MatchOperand(taicpu(p).oper[0]^,taicpu(p).oper[1]^)
        then
          begin
            DebugMsg(SPeepholeOptimization + 'Mov2Nop 1 done',p);
            { take care of the register (de)allocs following p }
            RemoveCurrentP(p, hp1);
            Result:=true;
            exit;
          end;

        { All the next optimisations require a next instruction }
        if not GetNextInstruction_p or (hp1.typ <> ait_instruction) then
          Exit;

        { Look for:
            mov %reg1,%reg2
            ??? %reg2,r/m
          Change to:
            mov %reg1,%reg2
            ??? %reg1,r/m
        }
        if MatchOpType(taicpu(p), top_reg, top_reg) then
          begin
            CurrentReg := taicpu(p).oper[1]^.reg;

            if RegReadByInstruction(CurrentReg, hp1) and
              DeepMOVOpt(taicpu(p), taicpu(hp1)) then
              begin
                TransferUsedRegs(TmpUsedRegs);
                UpdateUsedRegs(TmpUsedRegs, tai(p.Next));

                if not RegUsedAfterInstruction(CurrentReg, hp1, TmpUsedRegs) and
                  { Just in case something didn't get modified (e.g. an
                    implicit register) }
                  not RegReadByInstruction(CurrentReg, hp1) then
                  begin
                    { We can remove the original MOV }
                    DebugMsg(SPeepholeOptimization + 'Mov2Nop 3 done',p);
                    RemoveCurrentp(p, hp1);

                    { TmpUsedRegs contains the results of "UpdateUsedRegs(tai(p.Next))" already,
                      so just restore it to UsedRegs instead of calculating it again }
                    RestoreUsedRegs(TmpUsedRegs);
                    Result := True;
                    Exit;
                  end;

                { If we know a MOV instruction has become a null operation, we might as well
                  get rid of it now to save time. }
                if (taicpu(hp1).opcode = A_MOV) and
                  (taicpu(hp1).oper[1]^.typ = top_reg) and
                  SuperRegistersEqual(taicpu(hp1).oper[1]^.reg, taicpu(p).oper[0]^.reg) and
                  { Just being a register is enough to confirm it's a null operation }
                  (taicpu(hp1).oper[0]^.typ = top_reg) then
                  begin

                    Result := True;

                    { Speed-up to reduce a pipeline stall... if we had something like...

                        movl %eax,%edx
                        movw %dx,%ax

                      ... the second instruction would change to movw %ax,%ax, but
                      given that it is now %ax that's active rather than %eax,
                      penalties might occur due to a partial register write, so instead,
                      change it to a MOVZX instruction when optimising for speed.
                    }
                    if not (cs_opt_size in current_settings.optimizerswitches) and
                      IsMOVZXAcceptable and
                      (taicpu(hp1).opsize < taicpu(p).opsize)
{$ifdef x86_64}
                      { operations already implicitly set the upper 64 bits to zero }
                      and not ((taicpu(hp1).opsize = S_L) and (taicpu(p).opsize = S_Q))
{$endif x86_64}
                      then
                      begin
                        CurrentReg := taicpu(hp1).oper[1]^.reg;

                        DebugMsg(SPeepholeOptimization + 'Zero-extension to minimise pipeline stall (Mov2Movz)',hp1);
                        case taicpu(p).opsize of
                          S_W:
                            if taicpu(hp1).opsize = S_B then
                              taicpu(hp1).opsize := S_BL
                            else
                              InternalError(2020012911);
                          S_L{$ifdef x86_64}, S_Q{$endif x86_64}:
                            case taicpu(hp1).opsize of
                              S_B:
                                taicpu(hp1).opsize := S_BL;
                              S_W:
                                taicpu(hp1).opsize := S_WL;
                              else
                                InternalError(2020012912);
                            end;
                          else
                            InternalError(2020012910);
                        end;

                        taicpu(hp1).opcode := A_MOVZX;
                        taicpu(hp1).oper[1]^.reg := newreg(getregtype(CurrentReg), getsupreg(CurrentReg), R_SUBD)
                      end
                    else
                      begin
                        GetNextInstruction_p := GetNextInstruction(hp1, hp2);
                        DebugMsg(SPeepholeOptimization + 'Mov2Nop 4 done',hp1);
                        RemoveInstruction(hp1);

                        { The instruction after what was hp1 is now the immediate next instruction,
                          so we can continue to make optimisations if it's present }
                        if not GetNextInstruction_p or (hp2.typ <> ait_instruction) then
                          Exit;

                        hp1 := hp2;
                      end;
                  end;

              end;
          end;

        { Depending on the DeepMOVOpt above, it may turn out that hp1 completely
          overwrites the original destination register.  e.g.

          movl   ###,%reg2d
          movslq ###,%reg2q (### doesn't have to be the same as the first one)

          In this case, we can remove the MOV (Go to "Mov2Nop 5" below)
        }
        if (taicpu(p).oper[1]^.typ = top_reg) and
          MatchInstruction(hp1, [A_LEA, A_MOV, A_MOVSX, A_MOVZX{$ifdef x86_64}, A_MOVSXD{$endif x86_64}], []) and
          (taicpu(hp1).oper[1]^.typ = top_reg) and
          Reg1WriteOverwritesReg2Entirely(taicpu(hp1).oper[1]^.reg, taicpu(p).oper[1]^.reg) then
            begin
              if RegInOp(taicpu(p).oper[1]^.reg, taicpu(hp1).oper[0]^) then
                begin
                  if (taicpu(hp1).oper[0]^.typ = top_reg) then
                    case taicpu(p).oper[0]^.typ of
                      top_const:
                        { We have something like:

                          movb   $x,   %regb
                          movzbl %regb,%regd

                          Change to:

                          movl   $x,   %regd
                        }
                        begin
                          case taicpu(hp1).opsize of
                            S_BW:
                              begin
                                convert_mov_value(A_MOVSX, $FF);
                                setsubreg(taicpu(p).oper[1]^.reg, R_SUBW);
                                taicpu(p).opsize := S_W;
                              end;
                            S_BL:
                              begin
                                convert_mov_value(A_MOVSX, $FF);
                                setsubreg(taicpu(p).oper[1]^.reg, R_SUBD);
                                taicpu(p).opsize := S_L;
                              end;
                            S_WL:
                              begin
                                convert_mov_value(A_MOVSX, $FFFF);
                                setsubreg(taicpu(p).oper[1]^.reg, R_SUBD);
                                taicpu(p).opsize := S_L;
                              end;
{$ifdef x86_64}
                            S_BQ:
                              begin
                                convert_mov_value(A_MOVSX, $FF);
                                setsubreg(taicpu(p).oper[1]^.reg, R_SUBQ);
                                taicpu(p).opsize := S_Q;
                              end;
                            S_WQ:
                              begin
                                convert_mov_value(A_MOVSX, $FFFF);
                                setsubreg(taicpu(p).oper[1]^.reg, R_SUBQ);
                                taicpu(p).opsize := S_Q;
                              end;
                            S_LQ:
                              begin
                                convert_mov_value(A_MOVSXD, $FFFFFFFF);  { Note it's MOVSXD, not MOVSX }
                                setsubreg(taicpu(p).oper[1]^.reg, R_SUBQ);
                                taicpu(p).opsize := S_Q;
                              end;
{$endif x86_64}
                            else
                              { If hp1 was a MOV instruction, it should have been
                                optimised already }
                              InternalError(2020021001);
                          end;
                          DebugMsg(SPeepholeOptimization + 'MovMovXX2MovXX 2 done',p);
                          RemoveInstruction(hp1);
                          Result := True;
                          Exit;
                        end;
                      top_ref:
                        { We have something like:

                          movb   mem,  %regb
                          movzbl %regb,%regd

                          Change to:

                          movzbl mem,  %regd
                        }
                        if (taicpu(p).oper[0]^.ref^.refaddr<>addr_full) and (IsMOVZXAcceptable or (taicpu(hp1).opcode<>A_MOVZX)) then
                          begin
                            DebugMsg(SPeepholeOptimization + 'MovMovXX2MovXX 1 done',p);
                            taicpu(hp1).loadref(0,taicpu(p).oper[0]^.ref^);
                            RemoveCurrentP(p, hp1);
                            Result:=True;
                            Exit;
                          end;
                      else
                        if (taicpu(hp1).opcode <> A_MOV) and (taicpu(hp1).opcode <> A_LEA) then
                          { Just to make a saving, since there are no more optimisations with MOVZX and MOVSX/D }
                          Exit;
                  end;
                end
             { The RegInOp check makes sure that movl r/m,%reg1l; movzbl (%reg1l),%reg1l"
               and "movl r/m,%reg1; leal $1(%reg1,%reg2),%reg1" etc. are not incorrectly
               optimised }
              else
                begin
                  DebugMsg(SPeepholeOptimization + 'Mov2Nop 5 done',p);
                  RemoveCurrentP(p, hp1);
                  Result := True;
                  Exit;
                end;
            end;

        if (taicpu(hp1).opcode = A_AND) and
          (taicpu(p).oper[1]^.typ = top_reg) and
          MatchOpType(taicpu(hp1),top_const,top_reg) then
          begin
            if MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[1]^) then
              begin
                case taicpu(p).opsize of
                  S_L:
                    if (taicpu(hp1).oper[0]^.val = $ffffffff) then
                      begin
                        { Optimize out:
                            mov x, %reg
                            and ffffffffh, %reg
                        }
                        DebugMsg(SPeepholeOptimization + 'MovAnd2Mov 1 done',p);
                        RemoveInstruction(hp1);
                        Result:=true;
                        exit;
                      end;
                  S_Q: { TODO: Confirm if this is even possible }
                    if (taicpu(hp1).oper[0]^.val = $ffffffffffffffff) then
                      begin
                        { Optimize out:
                            mov x, %reg
                            and ffffffffffffffffh, %reg
                        }
                        DebugMsg(SPeepholeOptimization + 'MovAnd2Mov 2 done',p);
                        RemoveInstruction(hp1);
                        Result:=true;
                        exit;
                      end;
                  else
                    ;
                end;
                if ((taicpu(p).oper[0]^.typ=top_reg) or
                  ((taicpu(p).oper[0]^.typ=top_ref) and (taicpu(p).oper[0]^.ref^.refaddr<>addr_full))) and
                  GetNextInstruction(hp1,hp2) and
                  MatchInstruction(hp2,A_TEST,[taicpu(p).opsize]) and
                  MatchOperand(taicpu(hp1).oper[1]^,taicpu(hp2).oper[1]^) and
                  MatchOperand(taicpu(hp2).oper[0]^,taicpu(hp2).oper[1]^) and
                  GetNextInstruction(hp2,hp3) and
                  MatchInstruction(hp3,A_Jcc,A_Setcc,[S_NO]) and
                  (taicpu(hp3).condition in [C_E,C_NE]) then
                  begin
                    TransferUsedRegs(TmpUsedRegs);
                    UpdateUsedRegs(TmpUsedRegs, tai(p.Next));
                    UpdateUsedRegs(TmpUsedRegs, tai(hp1.Next));
                    if not(RegUsedAfterInstruction(taicpu(hp2).oper[1]^.reg, hp2, TmpUsedRegs)) then
                      begin
                        DebugMsg(SPeepholeOptimization + 'MovAndTest2Test done',p);
                        taicpu(hp1).loadoper(1,taicpu(p).oper[0]^);
                        taicpu(hp1).opcode:=A_TEST;
                        RemoveInstruction(hp2);
                        RemoveCurrentP(p, hp1);
                        Result:=true;
                        exit;
                      end;
                  end;
              end
            else if IsMOVZXAcceptable and
              (taicpu(p).oper[1]^.typ = top_reg) and (taicpu(hp1).oper[1]^.typ = top_reg) and
              (taicpu(p).oper[0]^.typ <> top_const) and { MOVZX only supports registers and memory, not immediates (use MOV for that!) }
              (getsupreg(taicpu(p).oper[1]^.reg) = getsupreg(taicpu(hp1).oper[1]^.reg))
              then
              begin
                InputVal := debug_operstr(taicpu(p).oper[0]^);
                MaskNum := debug_tostr(taicpu(hp1).oper[0]^.val);

                case taicpu(p).opsize of
                  S_B:
                    if (taicpu(hp1).oper[0]^.val = $ff) then
                      begin
                        { Convert:
                            movb x, %regl        movb x, %regl
                            andw ffh, %regw      andl ffh, %regd
                          To:
                            movzbw x, %regd      movzbl x, %regd

                          (Identical registers, just different sizes)
                        }
                        RegName1 := debug_regname(taicpu(p).oper[1]^.reg); { 8-bit register name }
                        RegName2 := debug_regname(taicpu(hp1).oper[1]^.reg); { 16/32-bit register name }

                        case taicpu(hp1).opsize of
                          S_W: NewSize := S_BW;
                          S_L: NewSize := S_BL;
{$ifdef x86_64}
                          S_Q: NewSize := S_BQ;
{$endif x86_64}
                          else
                            InternalError(2018011510);
                        end;
                      end
                    else
                      NewSize := S_NO;
                  S_W:
                    if (taicpu(hp1).oper[0]^.val = $ffff) then
                      begin
                        { Convert:
                            movw x, %regw
                            andl ffffh, %regd
                          To:
                            movzwl x, %regd

                          (Identical registers, just different sizes)
                        }
                        RegName1 := debug_regname(taicpu(p).oper[1]^.reg); { 16-bit register name }
                        RegName2 := debug_regname(taicpu(hp1).oper[1]^.reg); { 32-bit register name }

                        case taicpu(hp1).opsize of
                          S_L: NewSize := S_WL;
{$ifdef x86_64}
                          S_Q: NewSize := S_WQ;
{$endif x86_64}
                          else
                            InternalError(2018011511);
                        end;
                      end
                    else
                      NewSize := S_NO;
                  else
                    NewSize := S_NO;
                end;

                if NewSize <> S_NO then
                  begin
                    PreMessage := 'mov' + debug_opsize2str(taicpu(p).opsize) + ' ' + InputVal + ',' + RegName1;

                    { The actual optimization }
                    taicpu(p).opcode := A_MOVZX;
                    taicpu(p).changeopsize(NewSize);
                    taicpu(p).oper[1]^ := taicpu(hp1).oper[1]^;

                    { Safeguard if "and" is followed by a conditional command }
                    TransferUsedRegs(TmpUsedRegs);
                    UpdateUsedRegs(TmpUsedRegs,tai(p.next));

                    if (RegUsedAfterInstruction(NR_DEFAULTFLAGS, hp1, TmpUsedRegs)) then
                      begin
                        { At this point, the "and" command is effectively equivalent to
                          "test %reg,%reg". This will be handled separately by the
                          Peephole Optimizer. [Kit] }

                        DebugMsg(SPeepholeOptimization + PreMessage +
                          ' -> movz' + debug_opsize2str(NewSize) + ' ' + InputVal + ',' + RegName2, p);
                      end
                    else
                      begin
                        DebugMsg(SPeepholeOptimization + PreMessage + '; and' + debug_opsize2str(taicpu(hp1).opsize) + ' $' + MaskNum + ',' + RegName2 +
                          ' -> movz' + debug_opsize2str(NewSize) + ' ' + InputVal + ',' + RegName2, p);

                        RemoveInstruction(hp1);
                      end;

                    Result := True;
                    Exit;

                  end;
              end;
          end;
        { Next instruction is also a MOV ? }
        if MatchInstruction(hp1,A_MOV,[taicpu(p).opsize]) then
          begin
            if (taicpu(p).oper[1]^.typ = top_reg) and
              MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[0]^) then
              begin
                CurrentReg := taicpu(p).oper[1]^.reg;
                TransferUsedRegs(TmpUsedRegs);
                UpdateUsedRegs(TmpUsedRegs, tai(p.Next));
                { we have
                    mov x, %treg
                    mov %treg, y
                }
                if not(RegInOp(CurrentReg, taicpu(hp1).oper[1]^)) then
                  if not(RegUsedAfterInstruction(CurrentReg, hp1, TmpUsedRegs)) then
                    { we've got

                      mov x, %treg
                      mov %treg, y

                      with %treg is not used after }
                    case taicpu(p).oper[0]^.typ Of
                      { top_reg is covered by DeepMOVOpt }
                      top_const:
                        begin
                          { change
                              mov const, %treg
                              mov %treg, y

                              to

                              mov const, y
                          }
                          if (taicpu(hp1).oper[1]^.typ=top_reg) or
                            ((taicpu(p).oper[0]^.val>=low(longint)) and (taicpu(p).oper[0]^.val<=high(longint))) then
                            begin
                              if taicpu(hp1).oper[1]^.typ=top_reg then
                                AllocRegBetween(taicpu(hp1).oper[1]^.reg,p,hp1,usedregs);
                              taicpu(p).loadOper(1,taicpu(hp1).oper[1]^);
                              DebugMsg(SPeepholeOptimization + 'MovMov2Mov 5 done',p);
                              RemoveInstruction(hp1);
                              Result:=true;
                              Exit;
                            end;
                        end;
                      top_ref:
                        if (taicpu(hp1).oper[1]^.typ = top_reg) then
                          begin
                            { change
                                 mov mem, %treg
                                 mov %treg, %reg

                                 to

                                 mov mem, %reg"
                            }
                            taicpu(p).loadreg(1, taicpu(hp1).oper[1]^.reg);
                            DebugMsg(SPeepholeOptimization + 'MovMov2Mov 3 done',p);
                            RemoveInstruction(hp1);
                            Result:=true;
                            Exit;
                          end;
                      else
                        ;
                    end
                  else
                    { %treg is used afterwards, but all eventualities
                      other than the first MOV instruction being a constant
                      are covered by DeepMOVOpt, so only check for that }
                    if (taicpu(p).oper[0]^.typ = top_const) and
                      (
                        { For MOV operations, a size saving is only made if the register/const is byte-sized }
                        not (cs_opt_size in current_settings.optimizerswitches) or
                        (taicpu(hp1).opsize = S_B)
                      ) and
                      (
                        (taicpu(hp1).oper[1]^.typ = top_reg) or
                        ((taicpu(p).oper[0]^.val >= low(longint)) and (taicpu(p).oper[0]^.val <= high(longint)))
                      ) then
                      begin
                        DebugMsg(SPeepholeOptimization + debug_operstr(taicpu(hp1).oper[0]^) + ' = $' + debug_tostr(taicpu(p).oper[0]^.val) + '; changed to minimise pipeline stall (MovMov2Mov 6b)',hp1);
                        taicpu(hp1).loadconst(0, taicpu(p).oper[0]^.val);
                      end;
              end;
            if (taicpu(hp1).oper[0]^.typ = taicpu(p).oper[1]^.typ) and
               (taicpu(hp1).oper[1]^.typ = taicpu(p).oper[0]^.typ) then
                {  mov reg1, mem1     or     mov mem1, reg1
                   mov mem2, reg2            mov reg2, mem2}
              begin
                if OpsEqual(taicpu(hp1).oper[1]^,taicpu(p).oper[0]^) then
                  { mov reg1, mem1     or     mov mem1, reg1
                    mov mem2, reg1            mov reg2, mem1}
                  begin
                    if OpsEqual(taicpu(hp1).oper[0]^,taicpu(p).oper[1]^) then
                      { Removes the second statement from
                        mov reg1, mem1/reg2
                        mov mem1/reg2, reg1 }
                      begin
                        if taicpu(p).oper[0]^.typ=top_reg then
                          AllocRegBetween(taicpu(p).oper[0]^.reg,p,hp1,usedregs);
                        DebugMsg(SPeepholeOptimization + 'MovMov2Mov 1',p);
                        RemoveInstruction(hp1);
                        Result:=true;
                        exit;
                      end
                    else
                      begin
                        TransferUsedRegs(TmpUsedRegs);
                        UpdateUsedRegs(TmpUsedRegs, tai(hp1.next));
                        if (taicpu(p).oper[1]^.typ = top_ref) and
                          { mov reg1, mem1
                            mov mem2, reg1 }
                           (taicpu(hp1).oper[0]^.ref^.refaddr = addr_no) and
                           GetNextInstruction(hp1, hp2) and
                           MatchInstruction(hp2,A_CMP,[taicpu(p).opsize]) and
                           OpsEqual(taicpu(p).oper[1]^,taicpu(hp2).oper[0]^) and
                           OpsEqual(taicpu(p).oper[0]^,taicpu(hp2).oper[1]^) and
                           not(RegUsedAfterInstruction(taicpu(p).oper[0]^.reg, hp2, TmpUsedRegs)) then
                           { change                   to
                             mov reg1, mem1           mov reg1, mem1
                             mov mem2, reg1           cmp reg1, mem2
                             cmp mem1, reg1
                           }
                          begin
                            RemoveInstruction(hp2);
                            taicpu(hp1).opcode := A_CMP;
                            taicpu(hp1).loadref(1,taicpu(hp1).oper[0]^.ref^);
                            taicpu(hp1).loadreg(0,taicpu(p).oper[0]^.reg);
                            AllocRegBetween(taicpu(p).oper[0]^.reg,p,hp1,UsedRegs);
                            DebugMsg(SPeepholeOptimization + 'MovMovCmp2MovCmp done',hp1);
                          end;
                      end;
                  end
                else if (taicpu(p).oper[1]^.typ=top_ref) and
                  OpsEqual(taicpu(hp1).oper[0]^,taicpu(p).oper[1]^) then
                  begin
                    AllocRegBetween(taicpu(p).oper[0]^.reg,p,hp1,UsedRegs);
                    taicpu(hp1).loadreg(0,taicpu(p).oper[0]^.reg);
                    DebugMsg(SPeepholeOptimization + 'MovMov2MovMov1 done',p);
                  end
                else
                  begin
                    TransferUsedRegs(TmpUsedRegs);
                    if GetNextInstruction(hp1, hp2) and
                      MatchOpType(taicpu(p),top_ref,top_reg) and
                      MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[0]^) and
                      (taicpu(hp1).oper[1]^.typ = top_ref) and
                      MatchInstruction(hp2,A_MOV,[taicpu(p).opsize]) and
                      MatchOpType(taicpu(hp2),top_ref,top_reg) and
                      RefsEqual(taicpu(hp2).oper[0]^.ref^, taicpu(hp1).oper[1]^.ref^)  then
                      if not RegInRef(taicpu(hp2).oper[1]^.reg,taicpu(hp2).oper[0]^.ref^) and
                         not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp1,tmpUsedRegs)) then
                        {   mov mem1, %reg1
                            mov %reg1, mem2
                            mov mem2, reg2
                         to:
                            mov mem1, reg2
                            mov reg2, mem2}
                        begin
                          AllocRegBetween(taicpu(hp2).oper[1]^.reg,p,hp2,usedregs);
                          DebugMsg(SPeepholeOptimization + 'MovMovMov2MovMov 1 done',p);
                          taicpu(p).loadoper(1,taicpu(hp2).oper[1]^);
                          taicpu(hp1).loadoper(0,taicpu(hp2).oper[1]^);
                          RemoveInstruction(hp2);
                        end
{$ifdef i386}
                      { this is enabled for i386 only, as the rules to create the reg sets below
                        are too complicated for x86-64, so this makes this code too error prone
                        on x86-64
                      }
                      else if (taicpu(p).oper[1]^.reg <> taicpu(hp2).oper[1]^.reg) and
                        not(RegInRef(taicpu(p).oper[1]^.reg,taicpu(p).oper[0]^.ref^)) and
                        not(RegInRef(taicpu(hp2).oper[1]^.reg,taicpu(hp2).oper[0]^.ref^)) then
                        {   mov mem1, reg1         mov mem1, reg1
                            mov reg1, mem2         mov reg1, mem2
                            mov mem2, reg2         mov mem2, reg1
                         to:                    to:
                            mov mem1, reg1         mov mem1, reg1
                            mov mem1, reg2         mov reg1, mem2
                            mov reg1, mem2

                         or (if mem1 depends on reg1
                      and/or if mem2 depends on reg2)
                         to:
                             mov mem1, reg1
                             mov reg1, mem2
                             mov reg1, reg2
                        }
                        begin
                          taicpu(hp1).loadRef(0,taicpu(p).oper[0]^.ref^);
                          taicpu(hp1).loadReg(1,taicpu(hp2).oper[1]^.reg);
                          taicpu(hp2).loadRef(1,taicpu(hp2).oper[0]^.ref^);
                          taicpu(hp2).loadReg(0,taicpu(p).oper[1]^.reg);
                          AllocRegBetween(taicpu(p).oper[1]^.reg,p,hp2,usedregs);
                          if (taicpu(p).oper[0]^.ref^.base <> NR_NO) and
                             (getsupreg(taicpu(p).oper[0]^.ref^.base) in [RS_EAX,RS_EBX,RS_ECX,RS_EDX,RS_ESI,RS_EDI]) then
                            AllocRegBetween(taicpu(p).oper[0]^.ref^.base,p,hp2,usedregs);
                          if (taicpu(p).oper[0]^.ref^.index <> NR_NO) and
                             (getsupreg(taicpu(p).oper[0]^.ref^.index) in [RS_EAX,RS_EBX,RS_ECX,RS_EDX,RS_ESI,RS_EDI]) then
                            AllocRegBetween(taicpu(p).oper[0]^.ref^.index,p,hp2,usedregs);
                        end
                      else if (taicpu(hp1).Oper[0]^.reg <> taicpu(hp2).Oper[1]^.reg) then
                        begin
                          taicpu(hp2).loadReg(0,taicpu(hp1).Oper[0]^.reg);
                          AllocRegBetween(taicpu(p).oper[1]^.reg,p,hp2,usedregs);
                        end
                      else
                        begin
                          RemoveInstruction(hp2);
                        end
{$endif i386}
                        ;
                  end;
              end
            { movl [mem1],reg1
              movl [mem1],reg2

              to

              movl [mem1],reg1
              movl reg1,reg2
             }
             else if MatchOpType(taicpu(p),top_ref,top_reg) and
               MatchOpType(taicpu(hp1),top_ref,top_reg) and
               (taicpu(p).opsize = taicpu(hp1).opsize) and
               RefsEqual(taicpu(p).oper[0]^.ref^,taicpu(hp1).oper[0]^.ref^) and
               (taicpu(p).oper[0]^.ref^.volatility=[]) and
               (taicpu(hp1).oper[0]^.ref^.volatility=[]) and
               not(SuperRegistersEqual(taicpu(p).oper[1]^.reg,taicpu(hp1).oper[0]^.ref^.base)) and
               not(SuperRegistersEqual(taicpu(p).oper[1]^.reg,taicpu(hp1).oper[0]^.ref^.index)) then
               begin
                 DebugMsg(SPeepholeOptimization + 'MovMov2MovMov 2',p);
                 taicpu(hp1).loadReg(0,taicpu(p).oper[1]^.reg);
               end;

            {   movl const1,[mem1]
                movl [mem1],reg1

                to

                movl const1,reg1
                movl reg1,[mem1]
            }
            if MatchOpType(Taicpu(p),top_const,top_ref) and
                 MatchOpType(Taicpu(hp1),top_ref,top_reg) and
                 (taicpu(p).opsize = taicpu(hp1).opsize) and
                 RefsEqual(taicpu(hp1).oper[0]^.ref^,taicpu(p).oper[1]^.ref^) and
                 not(RegInRef(taicpu(hp1).oper[1]^.reg,taicpu(hp1).oper[0]^.ref^)) then
              begin
                AllocRegBetween(taicpu(hp1).oper[1]^.reg,p,hp1,usedregs);
                taicpu(hp1).loadReg(0,taicpu(hp1).oper[1]^.reg);
                taicpu(hp1).loadRef(1,taicpu(p).oper[1]^.ref^);
                taicpu(p).loadReg(1,taicpu(hp1).oper[0]^.reg);
                taicpu(hp1).fileinfo := taicpu(p).fileinfo;
                DebugMsg(SPeepholeOptimization + 'MovMov2MovMov 1',p);
                Result:=true;
                exit;
              end;

              { mov x,reg1; mov y,reg1 -> mov y,reg1 is handled by the Mov2Nop 5 optimisation }
          end;

        { search further than the next instruction for a mov }
        if
          { check as much as possible before the expensive GetNextInstructionUsingReg call }
          (taicpu(p).oper[1]^.typ = top_reg) and
          (taicpu(p).oper[0]^.typ in [top_reg,top_const]) and
          not RegModifiedByInstruction(taicpu(p).oper[1]^.reg, hp1) and
          { we work with hp2 here, so hp1 can be still used later on when
            checking for GetNextInstruction_p }
          { GetNextInstructionUsingReg only searches one instruction ahead unless -O3 is specified }
          GetNextInstructionUsingReg(hp1,hp2,taicpu(p).oper[1]^.reg) and
          (hp2.typ=ait_instruction) then
          begin
            case taicpu(hp2).opcode of
              A_MOV:
                if MatchOperand(taicpu(hp2).oper[0]^,taicpu(p).oper[1]^.reg) and
                  ((taicpu(p).oper[0]^.typ=top_const) or
                   ((taicpu(p).oper[0]^.typ=top_reg) and
                    not(RegUsedBetween(taicpu(p).oper[0]^.reg, p, hp2))
                   )
                  ) then
                  begin
                    { we have
                        mov x, %treg
                        mov %treg, y
                    }

                    TransferUsedRegs(TmpUsedRegs);
                    TmpUsedRegs[R_INTREGISTER].Update(tai(p.Next));

                    { We don't need to call UpdateUsedRegs for every instruction between
                      p and hp2 because the register we're concerned about will not
                      become deallocated (otherwise GetNextInstructionUsingReg would
                      have stopped at an earlier instruction). [Kit] }

                    TempRegUsed :=
                      RegUsedAfterInstruction(taicpu(p).oper[1]^.reg, hp2, TmpUsedRegs) or
                      RegReadByInstruction(taicpu(p).oper[1]^.reg, hp1);

                    case taicpu(p).oper[0]^.typ Of
                      top_reg:
                        begin
                          { change
                              mov %reg, %treg
                              mov %treg, y

                              to

                              mov %reg, y
                          }
                          CurrentReg := taicpu(p).oper[0]^.reg; { Saves on a handful of pointer dereferences }
                          RegName1 := debug_regname(taicpu(hp2).oper[0]^.reg);
                          if taicpu(hp2).oper[1]^.reg = CurrentReg then
                            begin
                              { %reg = y - remove hp2 completely (doing it here instead of relying on
                                the "mov %reg,%reg" optimisation might cut down on a pass iteration) }

                              if TempRegUsed then
                                begin
                                  DebugMsg(SPeepholeOptimization + debug_regname(CurrentReg) + ' = ' + RegName1 + '; removed unnecessary instruction (MovMov2MovNop 6b}',hp2);
                                  AllocRegBetween(CurrentReg, p, hp2, UsedRegs);
                                  RemoveInstruction(hp2);
                                end
                              else
                                begin
                                  RemoveInstruction(hp2);

                                  { We can remove the original MOV too }
                                  DebugMsg(SPeepholeOptimization + 'MovMov2NopNop 6b done',p);
                                  RemoveCurrentP(p, hp1);
                                  Result:=true;
                                  Exit;
                                end;
                            end
                          else
                            begin
                              AllocRegBetween(CurrentReg, p, hp2, UsedRegs);
                              taicpu(hp2).loadReg(0, CurrentReg);
                              if TempRegUsed then
                                begin
                                  { Don't remove the first instruction if the temporary register is in use }
                                  DebugMsg(SPeepholeOptimization + RegName1 + ' = ' + debug_regname(CurrentReg) + '; changed to minimise pipeline stall (MovMov2Mov 6a}',hp2);

                                  { No need to set Result to True. If there's another instruction later on
                                    that can be optimised, it will be detected when the main Pass 1 loop
                                    reaches what is now hp2 and passes it through OptPass1MOV. [Kit] };
                                end
                              else
                                begin
                                  DebugMsg(SPeepholeOptimization + 'MovMov2Mov 6 done',p);
                                  RemoveCurrentP(p, hp1);
                                  Result:=true;
                                  Exit;
                                end;
                            end;
                        end;
                      top_const:
                        if not (cs_opt_size in current_settings.optimizerswitches) or (taicpu(hp2).opsize = S_B) then
                          begin
                            { change
                                mov const, %treg
                                mov %treg, y

                                to

                                mov const, y
                            }
                            if (taicpu(hp2).oper[1]^.typ=top_reg) or
                              ((taicpu(p).oper[0]^.val>=low(longint)) and (taicpu(p).oper[0]^.val<=high(longint))) then
                              begin
                                RegName1 := debug_regname(taicpu(hp2).oper[0]^.reg);
                                taicpu(hp2).loadOper(0,taicpu(p).oper[0]^);

                                if TempRegUsed then
                                  begin
                                    { Don't remove the first instruction if the temporary register is in use }
                                    DebugMsg(SPeepholeOptimization + RegName1 + ' = ' + debug_tostr(taicpu(p).oper[0]^.val) + '; changed to minimise pipeline stall (MovMov2Mov 7a)',hp2);

                                    { No need to set Result to True. If there's another instruction later on
                                      that can be optimised, it will be detected when the main Pass 1 loop
                                      reaches what is now hp2 and passes it through OptPass1MOV. [Kit] };
                                  end
                                else
                                  begin
                                    DebugMsg(SPeepholeOptimization + 'MovMov2Mov 7 done',p);
                                    RemoveCurrentP(p, hp1);
                                    Result:=true;
                                    Exit;
                                  end;
                              end;
                          end;
                        else
                          Internalerror(2019103001);
                      end;
                  end;
              A_MOVZX, A_MOVSX{$ifdef x86_64}, A_MOVSXD{$endif x86_64}:
                if MatchOpType(taicpu(hp2), top_reg, top_reg) and
                  MatchOperand(taicpu(hp2).oper[0]^, taicpu(p).oper[1]^.reg) and
                  SuperRegistersEqual(taicpu(hp2).oper[1]^.reg, taicpu(p).oper[1]^.reg) then
                  begin
                    {
                      Change from:
                        mov    ###, %reg
                        ...
                        movs/z %reg,%reg  (Same register, just different sizes)

                      To:
                        movs/z ###, %reg  (Longer version)
                        ...
                        (remove)
                    }
                    DebugMsg(SPeepholeOptimization + 'MovMovs/z2Mov/s/z done', p);
                    taicpu(p).oper[1]^.reg := taicpu(hp2).oper[1]^.reg;

                    { Keep the first instruction as mov if ### is a constant }
                    if taicpu(p).oper[0]^.typ = top_const then
                      taicpu(p).opsize := reg2opsize(taicpu(hp2).oper[1]^.reg)
                    else
                      begin
                        taicpu(p).opcode := taicpu(hp2).opcode;
                        taicpu(p).opsize := taicpu(hp2).opsize;
                      end;

                    DebugMsg(SPeepholeOptimization + 'Removed movs/z instruction and extended earlier write (MovMovs/z2Mov/s/z)', hp2);
                    AllocRegBetween(taicpu(hp2).oper[1]^.reg, p, hp2, UsedRegs);
                    RemoveInstruction(hp2);

                    Result := True;
                    Exit;
                  end;
              else
                ;
            end;
          end;

        if (aoc_MovAnd2Mov_3 in OptsToCheck) and
          (taicpu(p).oper[1]^.typ = top_reg) and
          (taicpu(p).opsize = S_L) and
          GetNextInstructionUsingRegTrackingUse(p,hp2,taicpu(p).oper[1]^.reg) and
          (taicpu(hp2).opcode = A_AND) and
          (MatchOpType(taicpu(hp2),top_const,top_reg) or
           (MatchOpType(taicpu(hp2),top_reg,top_reg) and
            MatchOperand(taicpu(hp2).oper[0]^,taicpu(hp2).oper[1]^))
           ) then
          begin
            if SuperRegistersEqual(taicpu(p).oper[1]^.reg,taicpu(hp2).oper[1]^.reg) then
              begin
                if ((taicpu(hp2).oper[0]^.typ=top_const) and (taicpu(hp2).oper[0]^.val = $ffffffff)) or
                  ((taicpu(hp2).oper[0]^.typ=top_reg) and (taicpu(hp2).opsize=S_L)) then
                  begin
                    { Optimize out:
                        mov x, %reg
                        and ffffffffh, %reg
                    }
                    DebugMsg(SPeepholeOptimization + 'MovAnd2Mov 3 done',p);
                    RemoveInstruction(hp2);
                    Result:=true;
                    exit;
                  end;
              end;
          end;

        { leave out the mov from "mov reg, x(%frame_pointer); leave/ret" (with
          x >= RetOffset) as it doesn't do anything (it writes either to a
          parameter or to the temporary storage room for the function
          result)
        }
        if IsExitCode(hp1) and
          (taicpu(p).oper[1]^.typ = top_ref) and
          (taicpu(p).oper[1]^.ref^.index = NR_NO) and
          (
            (
              (taicpu(p).oper[1]^.ref^.base = current_procinfo.FramePointer) and
              not (
                assigned(current_procinfo.procdef.funcretsym) and
                (taicpu(p).oper[1]^.ref^.offset <= tabstractnormalvarsym(current_procinfo.procdef.funcretsym).localloc.reference.offset)
              )
            ) or
            { Also discard writes to the stack that are below the base pointer,
              as this is temporary storage rather than a function result on the
              stack, say. }
            (
              (taicpu(p).oper[1]^.ref^.base = NR_STACK_POINTER_REG) and
              (taicpu(p).oper[1]^.ref^.offset < current_procinfo.final_localsize)
            )
          ) then
          begin
            RemoveCurrentp(p, hp1);
            DebugMsg(SPeepholeOptimization + 'removed deadstore before leave/ret',p);
            RemoveLastDeallocForFuncRes(p);
            Result:=true;
            exit;
          end;
        if MatchOpType(taicpu(p),top_reg,top_ref) and
          MatchInstruction(hp1,A_CMP,A_TEST,[taicpu(p).opsize]) and
          (taicpu(hp1).oper[1]^.typ = top_ref) and
           RefsEqual(taicpu(p).oper[1]^.ref^, taicpu(hp1).oper[1]^.ref^) then
          begin
            { change
                mov reg1, mem1
                test/cmp x, mem1

                to

                mov reg1, mem1
                test/cmp x, reg1
            }
            taicpu(hp1).loadreg(1,taicpu(p).oper[0]^.reg);
            DebugMsg(SPeepholeOptimization + 'MovTestCmp2MovTestCmp 1',hp1);
            AllocRegBetween(taicpu(p).oper[0]^.reg,p,hp1,usedregs);
            exit;
          end;

        if MatchInstruction(hp1,A_LEA,[S_L{$ifdef x86_64},S_Q{$endif x86_64}]) and
          { If the flags register is in use, don't change the instruction to an
            ADD otherwise this will scramble the flags. [Kit] }
          not RegInUsedRegs(NR_DEFAULTFLAGS, UsedRegs) then
          begin
            if MatchOpType(Taicpu(p),top_ref,top_reg) and
               ((MatchReference(Taicpu(hp1).oper[0]^.ref^,Taicpu(hp1).oper[1]^.reg,Taicpu(p).oper[1]^.reg) and
                 (Taicpu(hp1).oper[0]^.ref^.base<>Taicpu(p).oper[1]^.reg)
                ) or
                (MatchReference(Taicpu(hp1).oper[0]^.ref^,Taicpu(p).oper[1]^.reg,Taicpu(hp1).oper[1]^.reg) and
                 (Taicpu(hp1).oper[0]^.ref^.index<>Taicpu(p).oper[1]^.reg)
                )
               ) then
               { mov reg1,ref
                 lea reg2,[reg1,reg2]

                 to

                 add reg2,ref}
              begin
                TransferUsedRegs(TmpUsedRegs);
                { reg1 may not be used afterwards }
                if not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg, hp1, TmpUsedRegs)) then
                  begin
                    Taicpu(hp1).opcode:=A_ADD;
                    Taicpu(hp1).oper[0]^.ref^:=Taicpu(p).oper[0]^.ref^;
                    DebugMsg(SPeepholeOptimization + 'MovLea2Add done',hp1);
                    RemoveCurrentp(p, hp1);
                    result:=true;
                    exit;
                  end;
              end;

            { If the LEA instruction can be converted into an arithmetic instruction,
              it may be possible to then fold it in the next optimisation, otherwise
              there's nothing more that can be optimised here. }
            if not ConvertLEA(taicpu(hp1)) then
              Exit;

          end;

        if (taicpu(p).oper[1]^.typ = top_reg) and
          (hp1.typ = ait_instruction) and
          GetNextInstruction(hp1, hp2) and
          MatchInstruction(hp2,A_MOV,[]) and
          (SuperRegistersEqual(taicpu(hp2).oper[0]^.reg,taicpu(p).oper[1]^.reg)) and
          (
            IsFoldableArithOp(taicpu(hp1), taicpu(p).oper[1]^.reg)
{$ifdef x86_64}
            or
            (
              (taicpu(p).opsize=S_L) and (taicpu(hp1).opsize=S_Q) and (taicpu(hp2).opsize=S_L) and
              IsFoldableArithOp(taicpu(hp1), newreg(R_INTREGISTER,getsupreg(taicpu(p).oper[1]^.reg),R_SUBQ))
            )
{$endif x86_64}
          ) then
          begin
            if OpsEqual(taicpu(hp2).oper[1]^, taicpu(p).oper[0]^) and
              (taicpu(hp2).oper[0]^.typ=top_reg) then
              { change   movsX/movzX    reg/ref, reg2
                         add/sub/or/... reg3/$const, reg2
                         mov            reg2 reg/ref
                         dealloc        reg2
                to
                         add/sub/or/... reg3/$const, reg/ref      }
              begin
                TransferUsedRegs(TmpUsedRegs);
                UpdateUsedRegs(TmpUsedRegs, tai(p.next));
                UpdateUsedRegs(TmpUsedRegs, tai(hp1.next));
                If not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp2,TmpUsedRegs)) then
                  begin
                    { by example:
                        movswl  %si,%eax        movswl  %si,%eax      p
                        decl    %eax            addl    %edx,%eax     hp1
                        movw    %ax,%si         movw    %ax,%si       hp2
                      ->
                        movswl  %si,%eax        movswl  %si,%eax      p
                        decw    %eax            addw    %edx,%eax     hp1
                        movw    %ax,%si         movw    %ax,%si       hp2
                    }
                    DebugMsg(SPeepholeOptimization + 'MovOpMov2Op ('+
                          debug_op2str(taicpu(p).opcode)+debug_opsize2str(taicpu(p).opsize)+' '+
                          debug_op2str(taicpu(hp1).opcode)+debug_opsize2str(taicpu(hp1).opsize)+' '+
                          debug_op2str(taicpu(hp2).opcode)+debug_opsize2str(taicpu(hp2).opsize)+')',p);
                    taicpu(hp1).changeopsize(taicpu(hp2).opsize);
                    {
                      ->
                        movswl  %si,%eax        movswl  %si,%eax      p
                        decw    %si             addw    %dx,%si       hp1
                        movw    %ax,%si         movw    %ax,%si       hp2
                    }
                    case taicpu(hp1).ops of
                      1:
                        begin
                          taicpu(hp1).loadoper(0, taicpu(hp2).oper[1]^);
                          if taicpu(hp1).oper[0]^.typ=top_reg then
                            setsubreg(taicpu(hp1).oper[0]^.reg,getsubreg(taicpu(hp2).oper[0]^.reg));
                        end;
                      2:
                        begin
                          taicpu(hp1).loadoper(1, taicpu(hp2).oper[1]^);
                          if (taicpu(hp1).oper[0]^.typ=top_reg) and
                            (taicpu(hp1).opcode<>A_SHL) and
                            (taicpu(hp1).opcode<>A_SHR) and
                            (taicpu(hp1).opcode<>A_SAR) then
                            setsubreg(taicpu(hp1).oper[0]^.reg,getsubreg(taicpu(hp2).oper[0]^.reg));
                        end;
                      else
                        internalerror(2008042701);
                    end;
                    {
                      ->
                        decw    %si             addw    %dx,%si       p
                    }
                    RemoveInstruction(hp2);
                    RemoveCurrentP(p, hp1);
                    Result:=True;
                    Exit;
                  end;
              end;
            if MatchOpType(taicpu(hp2),top_reg,top_reg) and
              not(SuperRegistersEqual(taicpu(hp1).oper[0]^.reg,taicpu(hp2).oper[1]^.reg)) and
              ((topsize2memsize[taicpu(hp1).opsize]<= topsize2memsize[taicpu(hp2).opsize]) or
               { opsize matters for these opcodes, we could probably work around this, but it is not worth the effort }
               ((taicpu(hp1).opcode<>A_SHL) and (taicpu(hp1).opcode<>A_SHR) and (taicpu(hp1).opcode<>A_SAR))
              )
{$ifdef i386}
              { byte registers of esi, edi, ebp, esp are not available on i386 }
              and ((taicpu(hp2).opsize<>S_B) or not(getsupreg(taicpu(hp1).oper[0]^.reg) in [RS_ESI,RS_EDI,RS_EBP,RS_ESP]))
              and ((taicpu(hp2).opsize<>S_B) or not(getsupreg(taicpu(p).oper[0]^.reg) in [RS_ESI,RS_EDI,RS_EBP,RS_ESP]))
{$endif i386}
              then
              { change   movsX/movzX    reg/ref, reg2
                         add/sub/or/... regX/$const, reg2
                         mov            reg2, reg3
                         dealloc        reg2
                to
                         movsX/movzX    reg/ref, reg3
                         add/sub/or/... reg3/$const, reg3
              }
              begin
                TransferUsedRegs(TmpUsedRegs);
                UpdateUsedRegs(TmpUsedRegs, tai(p.next));
                UpdateUsedRegs(TmpUsedRegs, tai(hp1.next));
                If not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp2,TmpUsedRegs)) then
                  begin
                    { by example:
                        movswl  %si,%eax        movswl  %si,%eax      p
                        decl    %eax            addl    %edx,%eax     hp1
                        movw    %ax,%si         movw    %ax,%si       hp2
                      ->
                        movswl  %si,%eax        movswl  %si,%eax      p
                        decw    %eax            addw    %edx,%eax     hp1
                        movw    %ax,%si         movw    %ax,%si       hp2
                    }
                    DebugMsg(SPeepholeOptimization + 'MovOpMov2MovOp ('+
                          debug_op2str(taicpu(p).opcode)+debug_opsize2str(taicpu(p).opsize)+' '+
                          debug_op2str(taicpu(hp1).opcode)+debug_opsize2str(taicpu(hp1).opsize)+' '+
                          debug_op2str(taicpu(hp2).opcode)+debug_opsize2str(taicpu(hp2).opsize)+')',p);
                    { limit size of constants as well to avoid assembler errors, but
                      check opsize to avoid overflow when left shifting the 1 }
                    if (taicpu(p).oper[0]^.typ=top_const) and (topsize2memsize[taicpu(hp2).opsize]<=63) then
                      taicpu(p).oper[0]^.val:=taicpu(p).oper[0]^.val and ((qword(1) shl topsize2memsize[taicpu(hp2).opsize])-1);
                    taicpu(hp1).changeopsize(taicpu(hp2).opsize);
                    taicpu(p).changeopsize(taicpu(hp2).opsize);
                    if taicpu(p).oper[0]^.typ=top_reg then
                      setsubreg(taicpu(p).oper[0]^.reg,getsubreg(taicpu(hp2).oper[0]^.reg));
                    taicpu(p).loadoper(1, taicpu(hp2).oper[1]^);
                    AllocRegBetween(taicpu(p).oper[1]^.reg,p,hp1,usedregs);
                    {
                      ->
                        movswl  %si,%eax        movswl  %si,%eax      p
                        decw    %si             addw    %dx,%si       hp1
                        movw    %ax,%si         movw    %ax,%si       hp2
                    }
                    case taicpu(hp1).ops of
                      1:
                        begin
                          taicpu(hp1).loadoper(0, taicpu(hp2).oper[1]^);
                          if taicpu(hp1).oper[0]^.typ=top_reg then
                            setsubreg(taicpu(hp1).oper[0]^.reg,getsubreg(taicpu(hp2).oper[0]^.reg));
                        end;
                      2:
                        begin
                          taicpu(hp1).loadoper(1, taicpu(hp2).oper[1]^);
                          if (taicpu(hp1).oper[0]^.typ=top_reg) and
                            (taicpu(hp1).opcode<>A_SHL) and
                            (taicpu(hp1).opcode<>A_SHR) and
                            (taicpu(hp1).opcode<>A_SAR) then
                            setsubreg(taicpu(hp1).oper[0]^.reg,getsubreg(taicpu(hp2).oper[0]^.reg));
                        end;
                      else
                        internalerror(2018111801);
                    end;
                    {
                      ->
                        decw    %si             addw    %dx,%si       p
                    }
                    RemoveInstruction(hp2);
                  end;
              end;

          end;
        if MatchInstruction(hp1,A_BTS,A_BTR,[Taicpu(p).opsize]) and
          GetNextInstruction(hp1, hp2) and
          MatchInstruction(hp2,A_OR,[Taicpu(p).opsize]) and
          MatchOperand(Taicpu(p).oper[0]^,0) and
          (Taicpu(p).oper[1]^.typ = top_reg) and
          MatchOperand(Taicpu(p).oper[1]^,Taicpu(hp1).oper[1]^) and
          MatchOperand(Taicpu(p).oper[1]^,Taicpu(hp2).oper[1]^) then
          { mov reg1,0
            bts reg1,operand1             -->      mov reg1,operand2
            or  reg1,operand2                      bts reg1,operand1}
          begin
            Taicpu(hp2).opcode:=A_MOV;
            asml.remove(hp1);
            insertllitem(hp2,hp2.next,hp1);
            RemoveCurrentp(p, hp1);
            Result:=true;
            exit;
          end;
      end;


   function TX86AsmOptimizer.OptPass1MOVXX(var p : tai) : boolean;
      var
        hp1 : tai;
      begin
        Result:=false;
        if taicpu(p).ops <> 2 then
          exit;
        if GetNextInstruction(p,hp1) and
          MatchInstruction(hp1,taicpu(p).opcode,[taicpu(p).opsize]) and
          (taicpu(hp1).ops = 2) then
          begin
            if (taicpu(hp1).oper[0]^.typ = taicpu(p).oper[1]^.typ) and
               (taicpu(hp1).oper[1]^.typ = taicpu(p).oper[0]^.typ) then
                {  movXX reg1, mem1     or     movXX mem1, reg1
                   movXX mem2, reg2            movXX reg2, mem2}
              begin
                if OpsEqual(taicpu(hp1).oper[1]^,taicpu(p).oper[0]^) then
                  { movXX reg1, mem1     or     movXX mem1, reg1
                    movXX mem2, reg1            movXX reg2, mem1}
                  begin
                    if OpsEqual(taicpu(hp1).oper[0]^,taicpu(p).oper[1]^) then
                      begin
                        { Removes the second statement from
                          movXX reg1, mem1/reg2
                          movXX mem1/reg2, reg1
                        }
                        if taicpu(p).oper[0]^.typ=top_reg then
                          AllocRegBetween(taicpu(p).oper[0]^.reg,p,hp1,usedregs);
                        { Removes the second statement from
                          movXX mem1/reg1, reg2
                          movXX reg2, mem1/reg1
                        }
                        if (taicpu(p).oper[1]^.typ=top_reg) and
                          not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp1,UsedRegs)) then
                          begin
                            DebugMsg(SPeepholeOptimization + 'MovXXMovXX2Nop 1 done',p);
                            RemoveInstruction(hp1);
                            RemoveCurrentp(p); { p will now be equal to the instruction that follows what was hp1 }
                          end
                        else
                          begin
                            DebugMsg(SPeepholeOptimization + 'MovXXMovXX2MoVXX 1 done',p);
                            RemoveInstruction(hp1);
                          end;
                        Result:=true;
                        exit;
                      end
                end;
            end;
        end;
      end;


    function TX86AsmOptimizer.OptPass1OP(var p : tai) : boolean;
      var
        hp1 : tai;
      begin
        result:=false;
        { replace
            <Op>X    %mreg1,%mreg2  // Op in [ADD,MUL]
            MovX     %mreg2,%mreg1
            dealloc  %mreg2

            by
            <Op>X    %mreg2,%mreg1
          ?
        }
        if GetNextInstruction(p,hp1) and
          { we mix single and double opperations here because we assume that the compiler
            generates vmovapd only after double operations and vmovaps only after single operations }
          MatchInstruction(hp1,A_MOVAPD,A_MOVAPS,[S_NO]) and
          MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[0]^) and
          MatchOperand(taicpu(p).oper[0]^,taicpu(hp1).oper[1]^) and
          (taicpu(p).oper[0]^.typ=top_reg) then
          begin
            TransferUsedRegs(TmpUsedRegs);
            UpdateUsedRegs(TmpUsedRegs, tai(p.next));
            if not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp1,TmpUsedRegs)) then
              begin
                taicpu(p).loadoper(0,taicpu(hp1).oper[0]^);
                taicpu(p).loadoper(1,taicpu(hp1).oper[1]^);
                DebugMsg(SPeepholeOptimization + 'OpMov2Op done',p);
                RemoveInstruction(hp1);
                result:=true;
              end;
          end;
      end;


    function TX86AsmOptimizer.OptPass1LEA(var p : tai) : boolean;
      var
        hp1, hp2, hp3: tai;
        l : ASizeInt;
        ref: Integer;
        saveref: treference;
      begin
        Result:=false;
        { removes seg register prefixes from LEA operations, as they
          don't do anything}
        taicpu(p).oper[0]^.ref^.Segment:=NR_NO;
        { changes "lea (%reg1), %reg2" into "mov %reg1, %reg2" }
        if (taicpu(p).oper[0]^.ref^.base <> NR_NO) and
           (taicpu(p).oper[0]^.ref^.index = NR_NO) and
           { do not mess with leas acessing the stack pointer }
           (taicpu(p).oper[1]^.reg <> NR_STACK_POINTER_REG) and
           (not(Assigned(taicpu(p).oper[0]^.ref^.Symbol))) then
          begin
            if (taicpu(p).oper[0]^.ref^.offset = 0) then
              begin
                if (taicpu(p).oper[0]^.ref^.base <> taicpu(p).oper[1]^.reg) then
                  begin
                    hp1:=taicpu.op_reg_reg(A_MOV,taicpu(p).opsize,taicpu(p).oper[0]^.ref^.base,
                      taicpu(p).oper[1]^.reg);
                    InsertLLItem(p.previous,p.next, hp1);
                    DebugMsg(SPeepholeOptimization + 'Lea2Mov done',hp1);
                    p.free;
                    p:=hp1;
                  end
                else
                  begin
                    DebugMsg(SPeepholeOptimization + 'Lea2Nop done',p);
                    RemoveCurrentP(p);
                  end;
                Result:=true;
                exit;
              end
            else if (
              { continue to use lea to adjust the stack pointer,
                it is the recommended way, but only if not optimizing for size }
                (taicpu(p).oper[1]^.reg<>NR_STACK_POINTER_REG) or
                (cs_opt_size in current_settings.optimizerswitches)
              ) and
              { If the flags register is in use, don't change the instruction
                to an ADD otherwise this will scramble the flags. [Kit] }
              not RegInUsedRegs(NR_DEFAULTFLAGS, UsedRegs) and
              ConvertLEA(taicpu(p)) then
              begin
                Result:=true;
                exit;
              end;
          end;
        if GetNextInstruction(p,hp1) and
          MatchInstruction(hp1,A_MOV,[taicpu(p).opsize]) and
          MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[0]^) and
          MatchOpType(Taicpu(hp1),top_reg,top_reg) and
          (taicpu(p).oper[1]^.reg<>NR_STACK_POINTER_REG) then
          begin
            TransferUsedRegs(TmpUsedRegs);
            UpdateUsedRegs(TmpUsedRegs, tai(p.next));
            if not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp1,TmpUsedRegs)) then
              begin
                taicpu(p).loadoper(1,taicpu(hp1).oper[1]^);
                DebugMsg(SPeepholeOptimization + 'LeaMov2Lea done',p);
                RemoveInstruction(hp1);
                result:=true;
              end;
          end;
        { changes
            lea offset1(regX), reg1
            lea offset2(reg1), reg1
            to
            lea offset1+offset2(regX), reg1 }

        { for now, we do not mess with the stack pointer, thought it might be usefull to remove
          unneeded lea sequences on the stack pointer, it needs to be tested in detail }
        if (taicpu(p).oper[1]^.reg <> NR_STACK_POINTER_REG) and
          GetNextInstructionUsingReg(p,hp1,taicpu(p).oper[1]^.reg) and
          MatchInstruction(hp1,A_LEA,[taicpu(p).opsize]) and
          MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[1]^) and
          (taicpu(p).oper[0]^.ref^.relsymbol=nil) and
          (taicpu(p).oper[0]^.ref^.segment=NR_NO) and
          (taicpu(p).oper[0]^.ref^.symbol=nil) and
          (((taicpu(hp1).oper[0]^.ref^.base=taicpu(p).oper[1]^.reg) and
            (taicpu(p).oper[0]^.ref^.scalefactor in [0,1]) and
            (taicpu(p).oper[0]^.ref^.index=NR_NO) and
            (taicpu(p).oper[0]^.ref^.index=taicpu(hp1).oper[0]^.ref^.index) and
            (taicpu(p).oper[0]^.ref^.scalefactor=taicpu(hp1).oper[0]^.ref^.scalefactor)
           ) or
           ((taicpu(hp1).oper[0]^.ref^.index=taicpu(p).oper[1]^.reg) and
            (taicpu(p).oper[0]^.ref^.index=NR_NO)
           ) or
           ((taicpu(hp1).oper[0]^.ref^.base=taicpu(p).oper[1]^.reg) and
            (taicpu(hp1).oper[0]^.ref^.scalefactor in [0,1]) and
            (taicpu(p).oper[0]^.ref^.base=NR_NO) and
            not(RegUsedBetween(taicpu(p).oper[0]^.ref^.index,p,hp1)))
          ) and
          not(RegUsedBetween(taicpu(p).oper[0]^.ref^.base,p,hp1)) and
          (taicpu(p).oper[0]^.ref^.relsymbol=taicpu(hp1).oper[0]^.ref^.relsymbol) and
          (taicpu(p).oper[0]^.ref^.segment=taicpu(hp1).oper[0]^.ref^.segment) and
          (taicpu(p).oper[0]^.ref^.symbol=taicpu(hp1).oper[0]^.ref^.symbol) then
          begin
            DebugMsg(SPeepholeOptimization + 'LeaLea2Lea done',p);
            if taicpu(hp1).oper[0]^.ref^.index=taicpu(p).oper[1]^.reg then
              begin
                taicpu(hp1).oper[0]^.ref^.index:=taicpu(p).oper[0]^.ref^.base;
                inc(taicpu(hp1).oper[0]^.ref^.offset,taicpu(p).oper[0]^.ref^.offset*max(taicpu(hp1).oper[0]^.ref^.scalefactor,1));
                { if the register is used as index and base, we have to increase for base as well
                  and adapt base }
                if taicpu(hp1).oper[0]^.ref^.base=taicpu(p).oper[1]^.reg then
                  begin
                    taicpu(hp1).oper[0]^.ref^.base:=taicpu(p).oper[0]^.ref^.base;
                    inc(taicpu(hp1).oper[0]^.ref^.offset,taicpu(p).oper[0]^.ref^.offset);
                  end;
              end
            else
              begin
                inc(taicpu(hp1).oper[0]^.ref^.offset,taicpu(p).oper[0]^.ref^.offset);
                taicpu(hp1).oper[0]^.ref^.base:=taicpu(p).oper[0]^.ref^.base;
              end;
            if taicpu(p).oper[0]^.ref^.index<>NR_NO then
              begin
                taicpu(hp1).oper[0]^.ref^.base:=taicpu(hp1).oper[0]^.ref^.index;
                taicpu(hp1).oper[0]^.ref^.index:=taicpu(p).oper[0]^.ref^.index;
                taicpu(hp1).oper[0]^.ref^.scalefactor:=taicpu(p).oper[0]^.ref^.scalefactor;
              end;
            RemoveCurrentP(p);
            result:=true;
            exit;
          end;
        { changes
            lea <ref1>, reg1
            <op> ...,<ref. with reg1>,...
            to
            <op> ...,<ref1>,... }
        if (taicpu(p).oper[1]^.reg<>current_procinfo.framepointer) and
          (taicpu(p).oper[1]^.reg<>NR_STACK_POINTER_REG) and
          GetNextInstruction(p,hp1) and
          (hp1.typ=ait_instruction) and
          not(MatchInstruction(hp1,A_LEA,[])) then
          begin
            { find a reference which uses reg1 }
            if (taicpu(hp1).ops>=1) and (taicpu(hp1).oper[0]^.typ=top_ref) and RegInOp(taicpu(p).oper[1]^.reg,taicpu(hp1).oper[0]^) then
              ref:=0
            else if (taicpu(hp1).ops>=2) and (taicpu(hp1).oper[1]^.typ=top_ref) and RegInOp(taicpu(p).oper[1]^.reg,taicpu(hp1).oper[1]^) then
              ref:=1
            else
              ref:=-1;
            if (ref<>-1) and
              { reg1 must be either the base or the index }
              ((taicpu(hp1).oper[ref]^.ref^.base=taicpu(p).oper[1]^.reg) xor (taicpu(hp1).oper[ref]^.ref^.index=taicpu(p).oper[1]^.reg)) then
              begin
                { reg1 can be removed from the reference }
                saveref:=taicpu(hp1).oper[ref]^.ref^;
                if taicpu(hp1).oper[ref]^.ref^.base=taicpu(p).oper[1]^.reg then
                  taicpu(hp1).oper[ref]^.ref^.base:=NR_NO
                else if taicpu(hp1).oper[ref]^.ref^.index=taicpu(p).oper[1]^.reg then
                  taicpu(hp1).oper[ref]^.ref^.index:=NR_NO
                else
                  Internalerror(2019111201);
                { check if the can insert all data of the lea into the second instruction }
                if ((taicpu(hp1).oper[ref]^.ref^.base=taicpu(p).oper[1]^.reg) or (taicpu(hp1).oper[ref]^.ref^.scalefactor in [0,1])) and
                  ((taicpu(p).oper[0]^.ref^.base=NR_NO) or (taicpu(hp1).oper[ref]^.ref^.base=NR_NO)) and
                  ((taicpu(p).oper[0]^.ref^.index=NR_NO) or (taicpu(hp1).oper[ref]^.ref^.index=NR_NO)) and
                  ((taicpu(p).oper[0]^.ref^.symbol=nil) or (taicpu(hp1).oper[ref]^.ref^.symbol=nil)) and
                  ((taicpu(p).oper[0]^.ref^.relsymbol=nil) or (taicpu(hp1).oper[ref]^.ref^.relsymbol=nil)) and
                  ((taicpu(p).oper[0]^.ref^.scalefactor in [0,1]) or (taicpu(hp1).oper[ref]^.ref^.scalefactor in [0,1])) and
                  (taicpu(p).oper[0]^.ref^.segment=NR_NO) and (taicpu(hp1).oper[ref]^.ref^.segment=NR_NO)
{$ifdef x86_64}
                  and (abs(taicpu(hp1).oper[ref]^.ref^.offset+taicpu(p).oper[0]^.ref^.offset)<=$7fffffff)
                  and (((taicpu(p).oper[0]^.ref^.base<>NR_RIP) and (taicpu(p).oper[0]^.ref^.index<>NR_RIP)) or
                       ((taicpu(hp1).oper[ref]^.ref^.base=NR_NO) and (taicpu(hp1).oper[ref]^.ref^.index=NR_NO))
                      )
{$endif x86_64}
                  then
                  begin
                    { reg1 might not used by the second instruction after it is remove from the reference }
                    if not(RegInInstruction(taicpu(p).oper[1]^.reg,taicpu(hp1))) then
                      begin
                        TransferUsedRegs(TmpUsedRegs);
                        UpdateUsedRegs(TmpUsedRegs, tai(p.next));
                        { reg1 is not updated so it might not be used afterwards }
                        if not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp1,TmpUsedRegs)) then
                          begin
                            DebugMsg(SPeepholeOptimization + 'LeaOp2Op done',p);
                            if taicpu(p).oper[0]^.ref^.base<>NR_NO then
                              taicpu(hp1).oper[ref]^.ref^.base:=taicpu(p).oper[0]^.ref^.base;
                            if taicpu(p).oper[0]^.ref^.index<>NR_NO then
                              taicpu(hp1).oper[ref]^.ref^.index:=taicpu(p).oper[0]^.ref^.index;
                            if taicpu(p).oper[0]^.ref^.symbol<>nil then
                              taicpu(hp1).oper[ref]^.ref^.symbol:=taicpu(p).oper[0]^.ref^.symbol;
                            if taicpu(p).oper[0]^.ref^.relsymbol<>nil then
                              taicpu(hp1).oper[ref]^.ref^.relsymbol:=taicpu(p).oper[0]^.ref^.relsymbol;
                            if not(taicpu(p).oper[0]^.ref^.scalefactor in [0,1]) then
                              taicpu(hp1).oper[ref]^.ref^.scalefactor:=taicpu(p).oper[0]^.ref^.scalefactor;
                            inc(taicpu(hp1).oper[ref]^.ref^.offset,taicpu(p).oper[0]^.ref^.offset);
                            RemoveCurrentP(p, hp1);
                            result:=true;
                            exit;
                          end
                      end;
                  end;
                { recover }
                taicpu(hp1).oper[ref]^.ref^:=saveref;
              end;
          end;
      end;


    function TX86AsmOptimizer.DoSubAddOpt(var p: tai): Boolean;
      var
        hp1 : tai;
      begin
        DoSubAddOpt := False;
        if GetLastInstruction(p, hp1) and
           (hp1.typ = ait_instruction) and
           (taicpu(hp1).opsize = taicpu(p).opsize) then
          case taicpu(hp1).opcode Of
            A_DEC:
              if (taicpu(hp1).oper[0]^.typ = top_reg) and
                MatchOperand(taicpu(hp1).oper[0]^,taicpu(p).oper[1]^) then
                begin
                  taicpu(p).loadConst(0,taicpu(p).oper[0]^.val+1);
                  RemoveInstruction(hp1);
                end;
             A_SUB:
               if MatchOpType(taicpu(hp1),top_const,top_reg) and
                 MatchOperand(taicpu(hp1).oper[1]^,taicpu(p).oper[1]^) then
                 begin
                   taicpu(p).loadConst(0,taicpu(p).oper[0]^.val+taicpu(hp1).oper[0]^.val);
                   RemoveInstruction(hp1);
                 end;
             A_ADD:
               begin
                 if MatchOpType(taicpu(hp1),top_const,top_reg) and
                   MatchOperand(taicpu(hp1).oper[1]^,taicpu(p).oper[1]^) then
                   begin
                     taicpu(p).loadConst(0,taicpu(p).oper[0]^.val-taicpu(hp1).oper[0]^.val);
                     RemoveInstruction(hp1);
                     if (taicpu(p).oper[0]^.val = 0) then
                       begin
                         hp1 := tai(p.next);
                         RemoveInstruction(p); { Note, the choice to not use RemoveCurrentp is deliberate }
                         if not GetLastInstruction(hp1, p) then
                           p := hp1;
                         DoSubAddOpt := True;
                       end
                   end;
               end;
             else
               ;
           end;
      end;


    function TX86AsmOptimizer.OptPass1Sub(var p : tai) : boolean;
{$ifdef i386}
      var
        hp1 : tai;
{$endif i386}
      begin
        Result:=false;
        { * change "subl $2, %esp; pushw x" to "pushl x"}
        { * change "sub/add const1, reg" or "dec reg" followed by
            "sub const2, reg" to one "sub ..., reg" }
        if MatchOpType(taicpu(p),top_const,top_reg) then
          begin
{$ifdef i386}
            if (taicpu(p).oper[0]^.val = 2) and
               (taicpu(p).oper[1]^.reg = NR_ESP) and
               { Don't do the sub/push optimization if the sub }
               { comes from setting up the stack frame (JM)    }
               (not(GetLastInstruction(p,hp1)) or
               not(MatchInstruction(hp1,A_MOV,[S_L]) and
                 MatchOperand(taicpu(hp1).oper[0]^,NR_ESP) and
                 MatchOperand(taicpu(hp1).oper[0]^,NR_EBP))) then
              begin
                hp1 := tai(p.next);
                while Assigned(hp1) and
                      (tai(hp1).typ in [ait_instruction]+SkipInstr) and
                      not RegReadByInstruction(NR_ESP,hp1) and
                      not RegModifiedByInstruction(NR_ESP,hp1) do
                  hp1 := tai(hp1.next);
                if Assigned(hp1) and
                  MatchInstruction(hp1,A_PUSH,[S_W]) then
                  begin
                    taicpu(hp1).changeopsize(S_L);
                    if taicpu(hp1).oper[0]^.typ=top_reg then
                      setsubreg(taicpu(hp1).oper[0]^.reg,R_SUBWHOLE);
                    hp1 := tai(p.next);
                    RemoveCurrentp(p, hp1);
                    Result:=true;
                    exit;
                  end;
              end;
{$endif i386}
            if DoSubAddOpt(p) then
              Result:=true;
          end;
      end;


    function TX86AsmOptimizer.OptPass1SHLSAL(var p : tai) : boolean;
      var
        TmpBool1,TmpBool2 : Boolean;
        tmpref : treference;
        hp1,hp2: tai;
        mask:    tcgint;
      begin
        Result:=false;

        { All these optimisations work on "shl/sal const,%reg" }
        if not MatchOpType(taicpu(p),top_const,top_reg) then
          Exit;

        if (taicpu(p).opsize in [S_L{$ifdef x86_64},S_Q{$endif x86_64}]) and
           (taicpu(p).oper[0]^.val <= 3) then
          { Changes "shl const, %reg32; add const/reg, %reg32" to one lea statement }
          begin
            { should we check the next instruction? }
            TmpBool1 := True;
            { have we found an add/sub which could be
              integrated in the lea? }
            TmpBool2 := False;
            reference_reset(tmpref,2,[]);
            TmpRef.index := taicpu(p).oper[1]^.reg;
            TmpRef.scalefactor := 1 shl taicpu(p).oper[0]^.val;
            while TmpBool1 and
                  GetNextInstruction(p, hp1) and
                  (tai(hp1).typ = ait_instruction) and
                  ((((taicpu(hp1).opcode = A_ADD) or
                     (taicpu(hp1).opcode = A_SUB)) and
                    (taicpu(hp1).oper[1]^.typ = Top_Reg) and
                    (taicpu(hp1).oper[1]^.reg = taicpu(p).oper[1]^.reg)) or
                   (((taicpu(hp1).opcode = A_INC) or
                     (taicpu(hp1).opcode = A_DEC)) and
                    (taicpu(hp1).oper[0]^.typ = Top_Reg) and
                    (taicpu(hp1).oper[0]^.reg = taicpu(p).oper[1]^.reg)) or
                    ((taicpu(hp1).opcode = A_LEA) and
                    (taicpu(hp1).oper[0]^.ref^.index = taicpu(p).oper[1]^.reg) and
                    (taicpu(hp1).oper[1]^.reg = taicpu(p).oper[1]^.reg))) and
                  (not GetNextInstruction(hp1,hp2) or
                   not instrReadsFlags(hp2)) Do
              begin
                TmpBool1 := False;
                if taicpu(hp1).opcode=A_LEA then
                  begin
                    if (TmpRef.base = NR_NO) and
                       (taicpu(hp1).oper[0]^.ref^.symbol=nil) and
                       (taicpu(hp1).oper[0]^.ref^.relsymbol=nil) and
                       (taicpu(hp1).oper[0]^.ref^.segment=NR_NO) and
                       ((taicpu(hp1).oper[0]^.ref^.scalefactor=0) or
                       (taicpu(hp1).oper[0]^.ref^.scalefactor*tmpref.scalefactor<=8)) then
                      begin
                        TmpBool1 := True;
                        TmpBool2 := True;
                        inc(TmpRef.offset, taicpu(hp1).oper[0]^.ref^.offset);
                        if taicpu(hp1).oper[0]^.ref^.scalefactor<>0 then
                          tmpref.scalefactor:=tmpref.scalefactor*taicpu(hp1).oper[0]^.ref^.scalefactor;
                        TmpRef.base := taicpu(hp1).oper[0]^.ref^.base;
                        RemoveInstruction(hp1);
                      end
                  end
                else if (taicpu(hp1).oper[0]^.typ = Top_Const) then
                  begin
                    TmpBool1 := True;
                    TmpBool2 := True;
                    case taicpu(hp1).opcode of
                      A_ADD:
                        inc(TmpRef.offset, longint(taicpu(hp1).oper[0]^.val));
                      A_SUB:
                        dec(TmpRef.offset, longint(taicpu(hp1).oper[0]^.val));
                      else
                        internalerror(2019050536);
                    end;
                    RemoveInstruction(hp1);
                  end
                else
                  if (taicpu(hp1).oper[0]^.typ = Top_Reg) and
                     (((taicpu(hp1).opcode = A_ADD) and
                       (TmpRef.base = NR_NO)) or
                      (taicpu(hp1).opcode = A_INC) or
                      (taicpu(hp1).opcode = A_DEC)) then
                    begin
                      TmpBool1 := True;
                      TmpBool2 := True;
                      case taicpu(hp1).opcode of
                        A_ADD:
                          TmpRef.base := taicpu(hp1).oper[0]^.reg;
                        A_INC:
                          inc(TmpRef.offset);
                        A_DEC:
                          dec(TmpRef.offset);
                        else
                          internalerror(2019050535);
                      end;
                      RemoveInstruction(hp1);
                    end;
              end;
            if TmpBool2
{$ifndef x86_64}
               or
               ((current_settings.optimizecputype < cpu_Pentium2) and
               (taicpu(p).oper[0]^.val <= 3) and
               not(cs_opt_size in current_settings.optimizerswitches))
{$endif x86_64}
              then
              begin
                if not(TmpBool2) and
                    (taicpu(p).oper[0]^.val=1) then
                  begin
                    hp1:=taicpu.Op_reg_reg(A_ADD,taicpu(p).opsize,
                      taicpu(p).oper[1]^.reg, taicpu(p).oper[1]^.reg)
                  end
                else
                  hp1:=taicpu.op_ref_reg(A_LEA, taicpu(p).opsize, TmpRef,
                              taicpu(p).oper[1]^.reg);
                DebugMsg(SPeepholeOptimization + 'ShlAddLeaSubIncDec2Lea',p);
                InsertLLItem(p.previous, p.next, hp1);
                p.free;
                p := hp1;
              end;
          end
{$ifndef x86_64}
        else if (current_settings.optimizecputype < cpu_Pentium2) then
          begin
            { changes "shl $1, %reg" to "add %reg, %reg", which is the same on a 386,
              but faster on a 486, and Tairable in both U and V pipes on the Pentium
              (unlike shl, which is only Tairable in the U pipe) }
            if taicpu(p).oper[0]^.val=1 then
                begin
                  hp1 := taicpu.Op_reg_reg(A_ADD,taicpu(p).opsize,
                            taicpu(p).oper[1]^.reg, taicpu(p).oper[1]^.reg);
                  InsertLLItem(p.previous, p.next, hp1);
                  p.free;
                  p := hp1;
                end
           { changes "shl $2, %reg" to "lea (,%reg,4), %reg"
             "shl $3, %reg" to "lea (,%reg,8), %reg }
           else if (taicpu(p).opsize = S_L) and
                   (taicpu(p).oper[0]^.val<= 3) then
             begin
               reference_reset(tmpref,2,[]);
               TmpRef.index := taicpu(p).oper[1]^.reg;
               TmpRef.scalefactor := 1 shl taicpu(p).oper[0]^.val;
               hp1 := taicpu.Op_ref_reg(A_LEA,S_L,TmpRef, taicpu(p).oper[1]^.reg);
               InsertLLItem(p.previous, p.next, hp1);
               p.free;
               p := hp1;
             end;
          end
{$endif x86_64}
        else if
          GetNextInstruction(p, hp1) and (hp1.typ = ait_instruction) and MatchOpType(taicpu(hp1), top_const, top_reg) and
          (
            (
              MatchInstruction(hp1, A_AND, [taicpu(p).opsize]) and
              SetAndTest(hp1, hp2)
{$ifdef x86_64}
            ) or
            (
              MatchInstruction(hp1, A_MOV, [taicpu(p).opsize]) and
              GetNextInstruction(hp1, hp2) and
              MatchInstruction(hp2, A_AND, [taicpu(p).opsize]) and
              MatchOpType(taicpu(hp2), top_reg, top_reg) and
              (taicpu(hp1).oper[1]^.reg = taicpu(hp2).oper[0]^.reg)
{$endif x86_64}
            )
          ) and
          (taicpu(p).oper[1]^.reg = taicpu(hp2).oper[1]^.reg) then
          begin
            { Change:
                shl x, %reg1
                mov -(1<<x), %reg2
                and %reg2, %reg1

              Or:
                shl x, %reg1
                and -(1<<x), %reg1

              To just:
                shl x, %reg1

              Since the and operation only zeroes bits that are already zero from the shl operation
            }
            case taicpu(p).oper[0]^.val of
               8:
                 mask:=$FFFFFFFFFFFFFF00;
               16:
                 mask:=$FFFFFFFFFFFF0000;
               32:
                 mask:=$FFFFFFFF00000000;
               63:
                 { Constant pre-calculated to prevent overflow errors with Int64 }
                 mask:=$8000000000000000;
               else
                 begin
                   if taicpu(p).oper[0]^.val >= 64 then
                     { Shouldn't happen realistically, since the register
                       is guaranteed to be set to zero at this point }
                     mask := 0
                   else
                     mask := -(Int64(1 shl taicpu(p).oper[0]^.val));
                 end;
            end;

            if taicpu(hp1).oper[0]^.val = mask then
              begin
                { Everything checks out, perform the optimisation, as long as
                  the FLAGS register isn't being used}
                TransferUsedRegs(TmpUsedRegs);
                UpdateUsedRegs(TmpUsedRegs, tai(p.next));

{$ifdef x86_64}
                if (hp1 <> hp2) then
                  begin
                    { "shl/mov/and" version }
                    UpdateUsedRegs(TmpUsedRegs, tai(hp1.next));

                    { Don't do the optimisation if the FLAGS register is in use }
                    if not(RegUsedAfterInstruction(NR_DEFAULTFLAGS, hp2, TmpUsedRegs)) then
                      begin
                        DebugMsg(SPeepholeOptimization + 'ShlMovAnd2Shl', p);
                        { Don't remove the 'mov' instruction if its register is used elsewhere }
                        if not(RegUsedAfterInstruction(taicpu(hp1).oper[1]^.reg, hp2, TmpUsedRegs)) then
                          begin
                            RemoveInstruction(hp1);
                            Result := True;
                          end;

                        { Only set Result to True if the 'mov' instruction was removed }
                        RemoveInstruction(hp2);
                      end;
                  end
                else
{$endif x86_64}
                  begin
                    { "shl/and" version }

                    { Don't do the optimisation if the FLAGS register is in use }
                    if not(RegUsedAfterInstruction(NR_DEFAULTFLAGS, hp1, TmpUsedRegs)) then
                      begin
                        DebugMsg(SPeepholeOptimization + 'ShlAnd2Shl', p);
                        RemoveInstruction(hp1);
                        Result := True;
                      end;
                  end;

                Exit;
              end
            else {$ifdef x86_64}if (hp1 = hp2) then{$endif x86_64}
              begin
                { Even if the mask doesn't allow for its removal, we might be
                  able to optimise the mask for the "shl/and" version, which
                  may permit other peephole optimisations }
{$ifdef DEBUG_AOPTCPU}
                mask := taicpu(hp1).oper[0]^.val and mask;
                if taicpu(hp1).oper[0]^.val <> mask then
                  begin
                    DebugMsg(
                      SPeepholeOptimization +
                      'Changed mask from $' + debug_tostr(taicpu(hp1).oper[0]^.val) +
                      ' to $' + debug_tostr(mask) +
                      'based on previous instruction (ShlAnd2ShlAnd)', hp1);
                    taicpu(hp1).oper[0]^.val := mask;
                  end;
{$else DEBUG_AOPTCPU}
                { If debugging is off, just set the operand even if it's the same }
                taicpu(hp1).oper[0]^.val := taicpu(hp1).oper[0]^.val and mask;
{$endif DEBUG_AOPTCPU}
              end;

          end;
      end;


    function TX86AsmOptimizer.OptPass1SETcc(var p: tai): boolean;
      var
        hp1,hp2,next: tai; SetC, JumpC: TAsmCond; Unconditional: Boolean;
      begin
        Result:=false;

        if MatchOpType(taicpu(p),top_reg) and
          GetNextInstruction(p, hp1) and
          ((MatchInstruction(hp1, A_TEST, [S_B]) and
           MatchOpType(taicpu(hp1),top_reg,top_reg) and
           (taicpu(hp1).oper[0]^.reg = taicpu(hp1).oper[1]^.reg)) or
           (MatchInstruction(hp1, A_CMP, [S_B]) and
            MatchOpType(taicpu(hp1),top_const,top_reg) and
            (taicpu(hp1).oper[0]^.val=0))
          ) and
          (taicpu(p).oper[0]^.reg = taicpu(hp1).oper[1]^.reg) and
          GetNextInstruction(hp1, hp2) and
          MatchInstruction(hp2, A_Jcc, []) then
          { Change from:             To:

            set(C) %reg              j(~C) label
            test   %reg,%reg/cmp $0,%reg
            je     label


            set(C) %reg              j(C)  label
            test   %reg,%reg/cmp $0,%reg
            jne    label
          }
          begin
            next := tai(p.Next);

            TransferUsedRegs(TmpUsedRegs);
            UpdateUsedRegs(TmpUsedRegs, next);
            UpdateUsedRegs(TmpUsedRegs, tai(hp1.next));

            JumpC := taicpu(hp2).condition;
            Unconditional := False;

            if conditions_equal(JumpC, C_E) then
              SetC := inverse_cond(taicpu(p).condition)
            else if conditions_equal(JumpC, C_NE) then
              SetC := taicpu(p).condition
            else
              { We've got something weird here (and inefficent) }
              begin
                DebugMsg('DEBUG: Inefficient jump - check code generation', p);
                SetC := C_NONE;

                { JAE/JNB will always branch (use 'condition_in', since C_AE <> C_NB normally) }
                if condition_in(C_AE, JumpC) then
                  Unconditional := True
                else
                  { Not sure what to do with this jump - drop out }
                  Exit;
              end;

            RemoveInstruction(hp1);

            if Unconditional then
              MakeUnconditional(taicpu(hp2))
            else
              begin
                if SetC = C_NONE then
                  InternalError(2018061401);

                taicpu(hp2).SetCondition(SetC);
              end;

            if not RegUsedAfterInstruction(taicpu(p).oper[0]^.reg, hp2, TmpUsedRegs) then
              begin
                RemoveCurrentp(p, hp2);
                Result := True;
              end;

            DebugMsg(SPeepholeOptimization + 'SETcc/TESTCmp/Jcc -> Jcc',p);
          end;
      end;


    function TX86AsmOptimizer.OptPass1FSTP(var p: tai): boolean;
      { returns true if a "continue" should be done after this optimization }
      var
        hp1, hp2: tai;
      begin
        Result := false;
        if MatchOpType(taicpu(p),top_ref) and
           GetNextInstruction(p, hp1) and
           (hp1.typ = ait_instruction) and
           (((taicpu(hp1).opcode = A_FLD) and
             (taicpu(p).opcode = A_FSTP)) or
            ((taicpu(p).opcode = A_FISTP) and
             (taicpu(hp1).opcode = A_FILD))) and
           MatchOpType(taicpu(hp1),top_ref) and
           (taicpu(hp1).opsize = taicpu(p).opsize) and
           RefsEqual(taicpu(p).oper[0]^.ref^, taicpu(hp1).oper[0]^.ref^) then
          begin
            { replacing fstp f;fld f by fst f is only valid for extended because of rounding }
            if (taicpu(p).opsize=S_FX) and
               GetNextInstruction(hp1, hp2) and
               (hp2.typ = ait_instruction) and
               IsExitCode(hp2) and
               (taicpu(p).oper[0]^.ref^.base = current_procinfo.FramePointer) and
               not(assigned(current_procinfo.procdef.funcretsym) and
                   (taicpu(p).oper[0]^.ref^.offset < tabstractnormalvarsym(current_procinfo.procdef.funcretsym).localloc.reference.offset)) and
               (taicpu(p).oper[0]^.ref^.index = NR_NO) then
              begin
                RemoveInstruction(hp1);
                RemoveCurrentP(p, hp2);
                RemoveLastDeallocForFuncRes(p);
                Result := true;
              end
            (* can't be done because the store operation rounds
            else
              { fst can't store an extended value! }
              if (taicpu(p).opsize <> S_FX) and
                 (taicpu(p).opsize <> S_IQ) then
                begin
                  if (taicpu(p).opcode = A_FSTP) then
                    taicpu(p).opcode := A_FST
                  else taicpu(p).opcode := A_FIST;
                  RemoveInstruction(hp1);
                end
            *)
          end;
      end;


     function TX86AsmOptimizer.OptPass1FLD(var p : tai) : boolean;
      var
       hp1, hp2: tai;
      begin
        result:=false;
        if MatchOpType(taicpu(p),top_reg) and
           GetNextInstruction(p, hp1) and
           (hp1.typ = Ait_Instruction) and
           MatchOpType(taicpu(hp1),top_reg,top_reg) and
           (taicpu(hp1).oper[0]^.reg = NR_ST) and
           (taicpu(hp1).oper[1]^.reg = NR_ST1) then
           { change                        to
               fld      reg               fxxx reg,st
               fxxxp    st, st1 (hp1)
             Remark: non commutative operations must be reversed!
           }
          begin
              case taicpu(hp1).opcode Of
                A_FMULP,A_FADDP,
                A_FSUBP,A_FDIVP,A_FSUBRP,A_FDIVRP:
                  begin
                    case taicpu(hp1).opcode Of
                      A_FADDP: taicpu(hp1).opcode := A_FADD;
                      A_FMULP: taicpu(hp1).opcode := A_FMUL;
                      A_FSUBP: taicpu(hp1).opcode := A_FSUBR;
                      A_FSUBRP: taicpu(hp1).opcode := A_FSUB;
                      A_FDIVP: taicpu(hp1).opcode := A_FDIVR;
                      A_FDIVRP: taicpu(hp1).opcode := A_FDIV;
                      else
                        internalerror(2019050534);
                    end;
                    taicpu(hp1).oper[0]^.reg := taicpu(p).oper[0]^.reg;
                    taicpu(hp1).oper[1]^.reg := NR_ST;
                    RemoveCurrentP(p, hp1);
                    Result:=true;
                    exit;
                  end;
                else
                  ;
              end;
          end
        else
          if MatchOpType(taicpu(p),top_ref) and
             GetNextInstruction(p, hp2) and
             (hp2.typ = Ait_Instruction) and
             MatchOpType(taicpu(hp2),top_reg,top_reg) and
             (taicpu(p).opsize in [S_FS, S_FL]) and
             (taicpu(hp2).oper[0]^.reg = NR_ST) and
             (taicpu(hp2).oper[1]^.reg = NR_ST1) then
            if GetLastInstruction(p, hp1) and
              MatchInstruction(hp1,A_FLD,A_FST,[taicpu(p).opsize]) and
              MatchOpType(taicpu(hp1),top_ref) and
              RefsEqual(taicpu(p).oper[0]^.ref^, taicpu(hp1).oper[0]^.ref^) then
              if ((taicpu(hp2).opcode = A_FMULP) or
                  (taicpu(hp2).opcode = A_FADDP)) then
              { change                      to
                  fld/fst   mem1  (hp1)       fld/fst   mem1
                  fld       mem1  (p)         fadd/
                  faddp/                       fmul     st, st
                  fmulp  st, st1 (hp2) }
                begin
                  RemoveCurrentP(p, hp1);
                  if (taicpu(hp2).opcode = A_FADDP) then
                    taicpu(hp2).opcode := A_FADD
                  else
                    taicpu(hp2).opcode := A_FMUL;
                  taicpu(hp2).oper[1]^.reg := NR_ST;
                end
              else
              { change              to
                  fld/fst mem1 (hp1)   fld/fst mem1
                  fld     mem1 (p)     fld      st}
                begin
                  taicpu(p).changeopsize(S_FL);
                  taicpu(p).loadreg(0,NR_ST);
                end
            else
              begin
                case taicpu(hp2).opcode Of
                  A_FMULP,A_FADDP,A_FSUBP,A_FDIVP,A_FSUBRP,A_FDIVRP:
            { change                        to
                fld/fst  mem1    (hp1)      fld/fst    mem1
                fld      mem2    (p)        fxxx       mem2
                fxxxp    st, st1 (hp2)                      }

                    begin
                      case taicpu(hp2).opcode Of
                        A_FADDP: taicpu(p).opcode := A_FADD;
                        A_FMULP: taicpu(p).opcode := A_FMUL;
                        A_FSUBP: taicpu(p).opcode := A_FSUBR;
                        A_FSUBRP: taicpu(p).opcode := A_FSUB;
                        A_FDIVP: taicpu(p).opcode := A_FDIVR;
                        A_FDIVRP: taicpu(p).opcode := A_FDIV;
                        else
                          internalerror(2019050533);
                      end;
                      RemoveInstruction(hp2);
                    end
                  else
                    ;
                end
              end
      end;


     function TX86AsmOptimizer.OptPass1Cmp(var p: tai): boolean;
       var
         v: TCGInt;
         hp1, hp2: tai;
       begin
         Result:=false;

         if taicpu(p).oper[0]^.typ = top_const then
           begin
             { Though GetNextInstruction can be factored out, it is an expensive
               call, so delay calling it until we have first checked cheaper
               conditions that are independent of it. }

             if (taicpu(p).oper[0]^.val = 0) and
               (taicpu(p).oper[1]^.typ = top_reg) and
               GetNextInstruction(p, hp1) and
               MatchInstruction(hp1,A_Jcc,A_SETcc,[]) then
               begin
                 hp2 := p;
                 { When dealing with "cmp $0,%reg", only ZF and SF contain
                   anything meaningful once it's converted to "test %reg,%reg";
                   additionally, some jumps will always (or never) branch, so
                   evaluate every jump immediately following the
                   comparison, optimising the conditions if possible.
                   Similarly with SETcc... those that are always set to 0 or 1
                   are changed to MOV instructions }
                 while GetNextInstruction(hp2, hp1) and
                   MatchInstruction(hp1,A_Jcc,A_SETcc,[]) do
                   begin
                     case taicpu(hp1).condition of
                       C_B, C_C, C_NAE, C_O:
                         { For B/NAE:
                             Will never branch since an unsigned integer can never be below zero
                           For C/O:
                             Result cannot overflow because 0 is being subtracted
                         }
                         begin
                           if taicpu(hp1).opcode = A_Jcc then
                             begin
                               DebugMsg(SPeepholeOptimization + 'Cmpcc2Testcc - condition B/C/NAE/O --> Never (jump removed)', hp1);
                               TAsmLabel(taicpu(hp1).oper[0]^.ref^.symbol).decrefs;
                               RemoveInstruction(hp1);
                               { Since hp1 was deleted, hp2 must not be updated }
                               Continue;
                             end
                           else
                             begin
                               DebugMsg(SPeepholeOptimization + 'Cmpcc2Testcc - condition B/C/NAE/O --> Never (set -> mov 0)', hp1);
                               { Convert "set(c) %reg" instruction to "movb 0,%reg" }
                               taicpu(hp1).opcode := A_MOV;
                               taicpu(hp1).ops := 2;
                               taicpu(hp1).condition := C_None;
                               taicpu(hp1).opsize := S_B;
                               taicpu(hp1).loadreg(1,taicpu(hp1).oper[0]^.reg);
                               taicpu(hp1).loadconst(0, 0);
                             end;
                         end;
                       C_BE, C_NA:
                         begin
                           { Will only branch if equal to zero }
                           DebugMsg(SPeepholeOptimization + 'Cmpcc2Testcc - condition BE/NA --> E', hp1);
                           taicpu(hp1).condition := C_E;
                         end;
                       C_A, C_NBE:
                         begin
                           { Will only branch if not equal to zero }
                           DebugMsg(SPeepholeOptimization + 'Cmpcc2Testcc - condition A/NBE --> NE', hp1);
                           taicpu(hp1).condition := C_NE;
                         end;
                       C_AE, C_NB, C_NC, C_NO:
                         begin
                           { Will always branch }
                           DebugMsg(SPeepholeOptimization + 'Cmpcc2Testcc - condition AE/NB/NC/NO --> Always', hp1);
                           if taicpu(hp1).opcode = A_Jcc then
                             begin
                               MakeUnconditional(taicpu(hp1));
                               { Any jumps/set that follow will now be dead code }
                               RemoveDeadCodeAfterJump(taicpu(hp1));
                               Break;
                             end
                           else
                             begin
                               { Convert "set(c) %reg" instruction to "movb 1,%reg" }
                               taicpu(hp1).opcode := A_MOV;
                               taicpu(hp1).ops := 2;
                               taicpu(hp1).condition := C_None;
                               taicpu(hp1).opsize := S_B;
                               taicpu(hp1).loadreg(1,taicpu(hp1).oper[0]^.reg);
                               taicpu(hp1).loadconst(0, 1);
                             end;
                         end;
                       C_None:
                         InternalError(2020012201);
                       C_P, C_PE, C_NP, C_PO:
                         { We can't handle parity checks and they should never be generated
                           after a general-purpose CMP (it's used in some floating-point
                           comparisons that don't use CMP) }
                         InternalError(2020012202);
                       else
                         { Zero/Equality, Sign, their complements and all of the
                           signed comparisons do not need to be converted };
                     end;
                     hp2 := hp1;
                   end;

                 { Convert the instruction to a TEST }

                 taicpu(p).opcode := A_TEST;
                 taicpu(p).loadreg(0,taicpu(p).oper[1]^.reg);
                 Result := True;
                 Exit;
               end
             else if (taicpu(p).oper[0]^.val = 1) and
               GetNextInstruction(p, hp1) and
               MatchInstruction(hp1,A_Jcc,A_SETcc,[]) and
               (taicpu(hp1).condition in [C_L, C_NGE]) then
               begin
                 { Convert;       To:
                     cmp $1,r/m     cmp $0,r/m
                     jl  @lbl       jle @lbl
                 }
                 DebugMsg(SPeepholeOptimization + 'Cmp1Jl2Cmp0Jle', p);
                 taicpu(p).oper[0]^.val := 0;
                 taicpu(hp1).condition := C_LE;

                 { If the instruction is now "cmp $0,%reg", convert it to a
                   TEST (and effectively do the work of the "cmp $0,%reg" in
                   the block above)

                   If it's a reference, we can get away with not setting
                   Result to True because he haven't evaluated the jump
                   in this pass yet.
                 }
                 if (taicpu(p).oper[1]^.typ = top_reg) then
                   begin
                     taicpu(p).opcode := A_TEST;
                     taicpu(p).loadreg(0,taicpu(p).oper[1]^.reg);
                     Result := True;
                   end;

                 Exit;
               end
             else if (taicpu(p).oper[1]^.typ = top_reg) then
               begin
                 { cmp register,$8000                neg register
                   je target                 -->     jo target

                   .... only if register is deallocated before jump.}
                 case Taicpu(p).opsize of
                   S_B: v:=$80;
                   S_W: v:=$8000;
                   S_L: v:=qword($80000000);
                   { S_Q will never happen: cmp with 64 bit constants is not possible }
                   S_Q:
                     Exit;
                   else
                     internalerror(2013112905);
                 end;

                 if (taicpu(p).oper[0]^.val=v) and
                    GetNextInstruction(p, hp1) and
                    MatchInstruction(hp1,A_Jcc,A_SETcc,[]) and
                    (Taicpu(hp1).condition in [C_E,C_NE]) then
                   begin
                     TransferUsedRegs(TmpUsedRegs);
                     UpdateUsedRegs(TmpUsedRegs,tai(p.next));
                     if not(RegInUsedRegs(Taicpu(p).oper[1]^.reg, TmpUsedRegs)) then
                       begin
                         DebugMsg(SPeepholeOptimization + 'CmpJe2NegJo done',p);
                         Taicpu(p).opcode:=A_NEG;
                         Taicpu(p).loadoper(0,Taicpu(p).oper[1]^);
                         Taicpu(p).clearop(1);
                         Taicpu(p).ops:=1;
                         if Taicpu(hp1).condition=C_E then
                           Taicpu(hp1).condition:=C_O
                         else
                           Taicpu(hp1).condition:=C_NO;
                         Result:=true;
                         exit;
                       end;
                   end;
               end;
           end;
     end;


   function TX86AsmOptimizer.OptPass1PXor(var p: tai): boolean;
     var
       hp1: tai;
     begin
       {
         remove the second (v)pxor from

           pxor reg,reg
           ...
           pxor reg,reg
       }
       Result:=false;
       if MatchOperand(taicpu(p).oper[0]^,taicpu(p).oper[1]^) and
         MatchOpType(taicpu(p),top_reg,top_reg) and
         GetNextInstructionUsingReg(p,hp1,taicpu(p).oper[0]^.reg) and
         MatchInstruction(taicpu(hp1),taicpu(p).opcode,[taicpu(p).opsize]) and
         MatchOperand(taicpu(p).oper[0]^,taicpu(hp1).oper[0]^) and
         MatchOperand(taicpu(hp1).oper[0]^,taicpu(hp1).oper[1]^) then
         begin
           DebugMsg(SPeepholeOptimization + 'PXorPXor2PXor done',hp1);
           RemoveInstruction(hp1);
           Result:=true;
           Exit;
         end;
     end;


   function TX86AsmOptimizer.OptPass1VPXor(var p: tai): boolean;
     var
       hp1: tai;
     begin
       {
         remove the second (v)pxor from

           (v)pxor reg,reg
           ...
           (v)pxor reg,reg
       }
       Result:=false;
       if MatchOperand(taicpu(p).oper[0]^,taicpu(p).oper[1]^,taicpu(p).oper[2]^) and
         MatchOpType(taicpu(p),top_reg,top_reg,top_reg) and
         GetNextInstructionUsingReg(p,hp1,taicpu(p).oper[0]^.reg) and
         MatchInstruction(taicpu(hp1),taicpu(p).opcode,[taicpu(p).opsize]) and
         MatchOperand(taicpu(p).oper[0]^,taicpu(hp1).oper[0]^) and
         MatchOperand(taicpu(hp1).oper[0]^,taicpu(hp1).oper[1]^,taicpu(hp1).oper[2]^) then
         begin
           DebugMsg(SPeepholeOptimization + 'VPXorVPXor2PXor done',hp1);
           RemoveInstruction(hp1);
           Result:=true;
           Exit;
         end;
     end;

   function TX86AsmOptimizer.OptPass1Imul(var p: tai): boolean;
     var
       hp1 : tai;
     begin
       result:=false;
       { replace
           IMul   const,%mreg1,%mreg2
           Mov    %reg2,%mreg3
           dealloc  %mreg3

           by
           Imul   const,%mreg1,%mreg23
       }
       if (taicpu(p).ops=3) and
         GetNextInstruction(p,hp1) and
         MatchInstruction(hp1,A_MOV,[taicpu(p).opsize]) and
         MatchOperand(taicpu(p).oper[2]^,taicpu(hp1).oper[0]^) and
         (taicpu(hp1).oper[1]^.typ=top_reg) then
         begin
           TransferUsedRegs(TmpUsedRegs);
           UpdateUsedRegs(TmpUsedRegs, tai(p.next));
           if not(RegUsedAfterInstruction(taicpu(hp1).oper[0]^.reg,hp1,TmpUsedRegs)) then
             begin
               taicpu(p).loadoper(2,taicpu(hp1).oper[1]^);
               DebugMsg(SPeepholeOptimization + 'ImulMov2Imul done',p);
               RemoveInstruction(hp1);
               result:=true;
             end;
         end;
     end;



   function TX86AsmOptimizer.OptPass2MOV(var p : tai) : boolean;

     function IsXCHGAcceptable: Boolean; inline;
       begin
         { Always accept if optimising for size }
         Result := (cs_opt_size in current_settings.optimizerswitches) or
           (
{$ifdef x86_64}
             { XCHG takes 3 cycles on AMD Athlon64 }
             (current_settings.optimizecputype >= cpu_core_i)
{$else x86_64}
             { From the Pentium M onwards, XCHG only has a latency of 2 rather
             than 3, so it becomes a saving compared to three MOVs with two of
             them able to execute simultaneously. [Kit] }
             (current_settings.optimizecputype >= cpu_PentiumM)
{$endif x86_64}
           );
       end;

      var
        NewRef: TReference;
       hp1,hp2,hp3: tai;
{$ifndef x86_64}
       hp4: tai;
       OperIdx: Integer;
{$endif x86_64}
      begin
        Result:=false;
        if not GetNextInstruction(p, hp1) then
          Exit;

        if MatchInstruction(hp1, A_JMP, [S_NO]) then
          begin
            { Sometimes the MOVs that OptPass2JMP produces can be improved
              further, but we can't just put this jump optimisation in pass 1
              because it tends to perform worse when conditional jumps are
              nearby (e.g. when converting CMOV instructions). [Kit] }
            if OptPass2JMP(hp1) then
              { call OptPass1MOV once to potentially merge any MOVs that were created }
              Result := OptPass1MOV(p)
              { OptPass2MOV will now exit but will be called again if OptPass1MOV
                returned True and the instruction is still a MOV, thus checking
                the optimisations below }

            { If OptPass2JMP returned False, no optimisations were done to
              the jump and there are no further optimisations that can be done
              to the MOV instruction on this pass }
          end
        else if MatchOpType(taicpu(p),top_reg,top_reg) and
          (taicpu(p).opsize in [S_L{$ifdef x86_64}, S_Q{$endif x86_64}]) and
          MatchInstruction(hp1,A_ADD,A_SUB,[taicpu(p).opsize]) and
          MatchOpType(taicpu(hp1),top_const,top_reg) and
          (taicpu(hp1).oper[1]^.reg = taicpu(p).oper[1]^.reg) and
          { be lazy, checking separately for sub would be slightly better }
          (abs(taicpu(hp1).oper[0]^.val)<=$7fffffff) then
          begin
            { Change:
                movl/q %reg1,%reg2      movl/q %reg1,%reg2
                addl/q $x,%reg2         subl/q $x,%reg2
              To:
                leal/q x(%reg1),%reg2   leal/q -x(%reg1),%reg2
            }
            TransferUsedRegs(TmpUsedRegs);
            UpdateUsedRegs(TmpUsedRegs, tai(p.Next));
            UpdateUsedRegs(TmpUsedRegs, tai(hp1.Next));
            if not GetNextInstruction(hp1, hp2) or
              (
              { The FLAGS register isn't always tracked properly, so do not
                perform this optimisation if a conditional statement follows }
                not RegReadByInstruction(NR_DEFAULTFLAGS, hp2) and
                not RegUsedAfterInstruction(NR_DEFAULTFLAGS, hp2, TmpUsedRegs)
              ) then
              begin
                reference_reset(NewRef, 1, []);
                NewRef.base := taicpu(p).oper[0]^.reg;
                NewRef.scalefactor := 1;

                if taicpu(hp1).opcode = A_ADD then
                  begin
                    DebugMsg(SPeepholeOptimization + 'MovAdd2Lea', p);
                    NewRef.offset := taicpu(hp1).oper[0]^.val;
                  end
                else
                  begin
                    DebugMsg(SPeepholeOptimization + 'MovSub2Lea', p);
                    NewRef.offset := -taicpu(hp1).oper[0]^.val;
                  end;

                taicpu(p).opcode := A_LEA;
                taicpu(p).loadref(0, NewRef);

                RemoveInstruction(hp1);

                Result := True;
                Exit;
              end;
          end
        else if MatchOpType(taicpu(p),top_reg,top_reg) and
{$ifdef x86_64}
          MatchInstruction(hp1,A_MOVZX,A_MOVSX,A_MOVSXD,[]) and
{$else x86_64}
          MatchInstruction(hp1,A_MOVZX,A_MOVSX,[]) and
{$endif x86_64}
          MatchOpType(taicpu(hp1),top_reg,top_reg) and
          (taicpu(hp1).oper[0]^.reg = taicpu(p).oper[1]^.reg) then
          { mov reg1, reg2                mov reg1, reg2
            movzx/sx reg2, reg3      to   movzx/sx reg1, reg3}
          begin
            taicpu(hp1).oper[0]^.reg := taicpu(p).oper[0]^.reg;
            DebugMsg(SPeepholeOptimization + 'mov %reg1,%reg2; movzx/sx %reg2,%reg3 -> mov %reg1,%reg2;movzx/sx %reg1,%reg3',p);

            { Don't remove the MOV command without first checking that reg2 isn't used afterwards,
              or unless supreg(reg3) = supreg(reg2)). [Kit] }

            TransferUsedRegs(TmpUsedRegs);
            UpdateUsedRegs(TmpUsedRegs, tai(p.next));

            if (getsupreg(taicpu(p).oper[1]^.reg) = getsupreg(taicpu(hp1).oper[1]^.reg)) or
              not RegUsedAfterInstruction(taicpu(p).oper[1]^.reg, hp1, TmpUsedRegs)
            then
              begin
                RemoveCurrentP(p, hp1);
                Result:=true;
              end;

            exit;
          end
        else if MatchOpType(taicpu(p),top_reg,top_reg) and
          IsXCHGAcceptable and
          { XCHG doesn't support 8-byte registers }
          (taicpu(p).opsize <> S_B) and
          MatchInstruction(hp1, A_MOV, []) and
          MatchOpType(taicpu(hp1),top_reg,top_reg) and
          (taicpu(hp1).oper[1]^.reg = taicpu(p).oper[0]^.reg) and
          GetNextInstruction(hp1, hp2) and
          MatchInstruction(hp2, A_MOV, []) and
          { Don't need to call MatchOpType for hp2 because the operand matches below cover for it }
          MatchOperand(taicpu(hp2).oper[0]^, taicpu(p).oper[1]^.reg) and
          MatchOperand(taicpu(hp2).oper[1]^, taicpu(hp1).oper[0]^.reg) then
          begin
            { mov %reg1,%reg2
              mov %reg3,%reg1        ->  xchg %reg3,%reg1
              mov %reg2,%reg3
              (%reg2 not used afterwards)

              Note that xchg takes 3 cycles to execute, and generally mov's take
              only one cycle apiece, but the first two mov's can be executed in
              parallel, only taking 2 cycles overall.  Older processors should
              therefore only optimise for size. [Kit]
            }
            TransferUsedRegs(TmpUsedRegs);
            UpdateUsedRegs(TmpUsedRegs, tai(p.Next));
            UpdateUsedRegs(TmpUsedRegs, tai(hp1.Next));

            if not RegUsedAfterInstruction(taicpu(p).oper[1]^.reg, hp2, TmpUsedRegs) then
              begin
                DebugMsg(SPeepholeOptimization + 'MovMovMov2XChg', p);
                AllocRegBetween(taicpu(hp2).oper[1]^.reg, p, hp1, UsedRegs);
                taicpu(hp1).opcode := A_XCHG;

                RemoveCurrentP(p, hp1);
                RemoveInstruction(hp2);

                Result := True;
                Exit;
              end;
          end
        else if MatchOpType(taicpu(p),top_reg,top_reg) and
          MatchInstruction(hp1, A_SAR, []) then
          begin
            if MatchOperand(taicpu(hp1).oper[0]^, 31) then
              begin
                { the use of %edx also covers the opsize being S_L }
                if MatchOperand(taicpu(hp1).oper[1]^, NR_EDX) then
                  begin
                    { Note it has to be specifically "movl %eax,%edx", and those specific sub-registers }
                    if (taicpu(p).oper[0]^.reg = NR_EAX) and
                      (taicpu(p).oper[1]^.reg = NR_EDX) then
                      begin
                        { Change:
                            movl %eax,%edx
                            sarl $31,%edx
                          To:
                            cltd
                        }
                        DebugMsg(SPeepholeOptimization + 'MovSar2Cltd', p);
                        RemoveInstruction(hp1);
                        taicpu(p).opcode := A_CDQ;
                        taicpu(p).opsize := S_NO;
                        taicpu(p).clearop(1);
                        taicpu(p).clearop(0);
                        taicpu(p).ops:=0;
                        Result := True;
                      end
                    else if (cs_opt_size in current_settings.optimizerswitches) and
                      (taicpu(p).oper[0]^.reg = NR_EDX) and
                      (taicpu(p).oper[1]^.reg = NR_EAX) then
                      begin
                        { Change:
                            movl %edx,%eax
                            sarl $31,%edx
                          To:
                            movl %edx,%eax
                            cltd

                          Note that this creates a dependency between the two instructions,
                            so only perform if optimising for size.
                        }
                        DebugMsg(SPeepholeOptimization + 'MovSar2MovCltd', p);
                        taicpu(hp1).opcode := A_CDQ;
                        taicpu(hp1).opsize := S_NO;
                        taicpu(hp1).clearop(1);
                        taicpu(hp1).clearop(0);
                        taicpu(hp1).ops:=0;
                      end;
{$ifndef x86_64}
                  end
                { Don't bother if CMOV is supported, because a more optimal
                  sequence would have been generated for the Abs() intrinsic }
                else if not(CPUX86_HAS_CMOV in cpu_capabilities[current_settings.cputype]) and
                  { the use of %eax also covers the opsize being S_L }
                  MatchOperand(taicpu(hp1).oper[1]^, NR_EAX) and
                  (taicpu(p).oper[0]^.reg = NR_EAX) and
                  (taicpu(p).oper[1]^.reg = NR_EDX) and
                  GetNextInstruction(hp1, hp2) and
                  MatchInstruction(hp2, A_XOR, [S_L]) and
                  MatchOperand(taicpu(hp2).oper[0]^, NR_EAX) and
                  MatchOperand(taicpu(hp2).oper[1]^, NR_EDX) and

                  GetNextInstruction(hp2, hp3) and
                  MatchInstruction(hp3, A_SUB, [S_L]) and
                  MatchOperand(taicpu(hp3).oper[0]^, NR_EAX) and
                  MatchOperand(taicpu(hp3).oper[1]^, NR_EDX) then
                  begin
                    { Change:
                        movl %eax,%edx
                        sarl $31,%eax
                        xorl %eax,%edx
                        subl %eax,%edx
                        (Instruction that uses %edx)
                        (%eax deallocated)
                        (%edx deallocated)
                      To:
                        cltd
                        xorl %edx,%eax  <-- Note the registers have swapped
                        subl %edx,%eax
                        (Instruction that uses %eax) <-- %eax rather than %edx
                    }

                    TransferUsedRegs(TmpUsedRegs);
                    UpdateUsedRegs(TmpUsedRegs, tai(p.Next));
                    UpdateUsedRegs(TmpUsedRegs, tai(hp1.Next));
                    UpdateUsedRegs(TmpUsedRegs, tai(hp2.Next));

                    if not RegUsedAfterInstruction(NR_EAX, hp3, TmpUsedRegs) then
                      begin
                        if GetNextInstruction(hp3, hp4) and
                          not RegModifiedByInstruction(NR_EDX, hp4) and
                          not RegUsedAfterInstruction(NR_EDX, hp4, TmpUsedRegs) then
                          begin
                            DebugMsg(SPeepholeOptimization + 'abs() intrinsic optimisation', p);

                            taicpu(p).opcode := A_CDQ;
                            taicpu(p).clearop(1);
                            taicpu(p).clearop(0);
                            taicpu(p).ops:=0;

                            RemoveInstruction(hp1);

                            taicpu(hp2).loadreg(0, NR_EDX);
                            taicpu(hp2).loadreg(1, NR_EAX);

                            taicpu(hp3).loadreg(0, NR_EDX);
                            taicpu(hp3).loadreg(1, NR_EAX);

                            AllocRegBetween(NR_EAX, hp3, hp4, TmpUsedRegs);
                            { Convert references in the following instruction (hp4) from %edx to %eax }
                            for OperIdx := 0 to taicpu(hp4).ops - 1 do
                              with taicpu(hp4).oper[OperIdx]^ do
                                case typ of
                                  top_reg:
                                    if getsupreg(reg) = RS_EDX then
                                      reg := newreg(R_INTREGISTER,RS_EAX,getsubreg(reg));
                                  top_ref:
                                    begin
                                      if getsupreg(reg) = RS_EDX then
                                        ref^.base := newreg(R_INTREGISTER,RS_EAX,getsubreg(reg));
                                      if getsupreg(reg) = RS_EDX then
                                        ref^.index := newreg(R_INTREGISTER,RS_EAX,getsubreg(reg));
                                    end;
                                  else
                                    ;
                                end;
                          end;
                      end;
{$else x86_64}
                  end;
              end
            else if MatchOperand(taicpu(hp1).oper[0]^, 63) and
              { the use of %rdx also covers the opsize being S_Q }
              MatchOperand(taicpu(hp1).oper[1]^, NR_RDX) then
              begin
                { Note it has to be specifically "movq %rax,%rdx", and those specific sub-registers }
                if (taicpu(p).oper[0]^.reg = NR_RAX) and
                  (taicpu(p).oper[1]^.reg = NR_RDX) then
                  begin
                    { Change:
                        movq %rax,%rdx
                        sarq $63,%rdx
                      To:
                        cqto
                    }
                    DebugMsg(SPeepholeOptimization + 'MovSar2Cqto', p);
                    RemoveInstruction(hp1);
                    taicpu(p).opcode := A_CQO;
                    taicpu(p).opsize := S_NO;
                    taicpu(p).clearop(1);
                    taicpu(p).clearop(0);
                    taicpu(p).ops:=0;
                    Result := True;
                  end
                else if (cs_opt_size in current_settings.optimizerswitches) and
                  (taicpu(p).oper[0]^.reg = NR_RDX) and
                  (taicpu(p).oper[1]^.reg = NR_RAX) then
                  begin
                    { Change:
                        movq %rdx,%rax
                        sarq $63,%rdx
                      To:
                        movq %rdx,%rax
                        cqto

                      Note that this creates a dependency between the two instructions,
                        so only perform if optimising for size.
                    }
                    DebugMsg(SPeepholeOptimization + 'MovSar2MovCqto', p);
                    taicpu(hp1).opcode := A_CQO;
                    taicpu(hp1).opsize := S_NO;
                    taicpu(hp1).clearop(1);
                    taicpu(hp1).clearop(0);
                    taicpu(hp1).ops:=0;
{$endif x86_64}
                  end;
              end;
          end
        else if MatchInstruction(hp1, A_MOV, []) and
          (taicpu(hp1).oper[1]^.typ = top_reg) then
          { Though "GetNextInstruction" could be factored out, along with
            the instructions that depend on hp2, it is an expensive call that
            should be delayed for as long as possible, hence we do cheaper
            checks first that are likely to be False. [Kit] }
          begin

            if MatchOperand(taicpu(p).oper[1]^, NR_EDX) and
              (
                (
                  (taicpu(hp1).oper[1]^.reg = NR_EAX) and
                  (
                    MatchOperand(taicpu(hp1).oper[0]^, taicpu(p).oper[0]^) or
                    MatchOperand(taicpu(hp1).oper[0]^, NR_EDX)
                  )
                ) or
                (
                  (taicpu(hp1).oper[1]^.reg = NR_EDX) and
                  (
                    MatchOperand(taicpu(hp1).oper[0]^, taicpu(p).oper[0]^) or
                    MatchOperand(taicpu(hp1).oper[0]^, NR_EAX)
                  )
                )
              ) and
              GetNextInstruction(hp1, hp2) and
              MatchInstruction(hp2, A_SAR, []) and
              MatchOperand(taicpu(hp2).oper[0]^, 31) then
              begin
                if MatchOperand(taicpu(hp2).oper[1]^, NR_EDX) then
                  begin
                    { Change:
                        movl r/m,%edx         movl r/m,%eax         movl r/m,%edx         movl r/m,%eax
                        movl %edx,%eax   or   movl %eax,%edx   or   movl r/m,%eax    or   movl r/m,%edx
                        sarl $31,%edx         sarl $31,%edx         sarl $31,%edx         sarl $31,%edx
                      To:
                        movl r/m,%eax    <- Note the change in register
                        cltd
                    }
                    DebugMsg(SPeepholeOptimization + 'MovMovSar2MovCltd', p);

                    AllocRegBetween(NR_EAX, p, hp1, UsedRegs);
                    taicpu(p).loadreg(1, NR_EAX);

                    taicpu(hp1).opcode := A_CDQ;
                    taicpu(hp1).clearop(1);
                    taicpu(hp1).clearop(0);
                    taicpu(hp1).ops:=0;

                    RemoveInstruction(hp2);
(*
{$ifdef x86_64}
                  end
                else if MatchOperand(taicpu(hp2).oper[1]^, NR_RDX) and
                  { This code sequence does not get generated - however it might become useful
                    if and when 128-bit signed integer types make an appearance, so the code
                    is kept here for when it is eventually needed. [Kit] }
                  (
                    (
                      (taicpu(hp1).oper[1]^.reg = NR_RAX) and
                      (
                        MatchOperand(taicpu(hp1).oper[0]^, taicpu(p).oper[0]^) or
                        MatchOperand(taicpu(hp1).oper[0]^, NR_RDX)
                      )
                    ) or
                    (
                      (taicpu(hp1).oper[1]^.reg = NR_RDX) and
                      (
                        MatchOperand(taicpu(hp1).oper[0]^, taicpu(p).oper[0]^) or
                        MatchOperand(taicpu(hp1).oper[0]^, NR_RAX)
                      )
                    )
                  ) and
                  GetNextInstruction(hp1, hp2) and
                  MatchInstruction(hp2, A_SAR, [S_Q]) and
                  MatchOperand(taicpu(hp2).oper[0]^, 63) and
                  MatchOperand(taicpu(hp2).oper[1]^, NR_RDX) then
                  begin
                    { Change:
                        movq r/m,%rdx         movq r/m,%rax         movq r/m,%rdx         movq r/m,%rax
                        movq %rdx,%rax   or   movq %rax,%rdx   or   movq r/m,%rax    or   movq r/m,%rdx
                        sarq $63,%rdx         sarq $63,%rdx         sarq $63,%rdx         sarq $63,%rdx
                      To:
                        movq r/m,%rax    <- Note the change in register
                        cqto
                    }
                    DebugMsg(SPeepholeOptimization + 'MovMovSar2MovCqto', p);

                    AllocRegBetween(NR_RAX, p, hp1, UsedRegs);
                    taicpu(p).loadreg(1, NR_RAX);

                    taicpu(hp1).opcode := A_CQO;
                    taicpu(hp1).clearop(1);
                    taicpu(hp1).clearop(0);
                    taicpu(hp1).ops:=0;

                    RemoveInstruction(hp2);
{$endif x86_64}
*)
                  end;
              end;
{$ifdef x86_64}
          end
        else if (taicpu(p).opsize = S_L) and
          (taicpu(p).oper[1]^.typ = top_reg) and
          (
            MatchInstruction(hp1, A_MOV,[]) and
            (taicpu(hp1).opsize = S_L) and
            (taicpu(hp1).oper[1]^.typ = top_reg)
          ) and (
            GetNextInstruction(hp1, hp2) and
            (tai(hp2).typ=ait_instruction) and
            (taicpu(hp2).opsize = S_Q) and
            (
              (
                MatchInstruction(hp2, A_ADD,[]) and
                (taicpu(hp2).opsize = S_Q) and
                (taicpu(hp2).oper[0]^.typ = top_reg) and (taicpu(hp2).oper[1]^.typ = top_reg) and
                (
                  (
                    (getsupreg(taicpu(hp2).oper[0]^.reg) = getsupreg(taicpu(p).oper[1]^.reg)) and
                    (getsupreg(taicpu(hp2).oper[1]^.reg) = getsupreg(taicpu(hp1).oper[1]^.reg))
                  ) or (
                    (getsupreg(taicpu(hp2).oper[0]^.reg) = getsupreg(taicpu(hp1).oper[1]^.reg)) and
                    (getsupreg(taicpu(hp2).oper[1]^.reg) = getsupreg(taicpu(p).oper[1]^.reg))
                  )
                )
              ) or (
                MatchInstruction(hp2, A_LEA,[]) and
                (taicpu(hp2).oper[0]^.ref^.offset = 0) and
                (taicpu(hp2).oper[0]^.ref^.scalefactor <= 1) and
                (
                  (
                    (getsupreg(taicpu(hp2).oper[0]^.ref^.base) = getsupreg(taicpu(p).oper[1]^.reg)) and
                    (getsupreg(taicpu(hp2).oper[0]^.ref^.index) = getsupreg(taicpu(hp1).oper[1]^.reg))
                  ) or (
                    (getsupreg(taicpu(hp2).oper[0]^.ref^.base) = getsupreg(taicpu(hp1).oper[1]^.reg)) and
                    (getsupreg(taicpu(hp2).oper[0]^.ref^.index) = getsupreg(taicpu(p).oper[1]^.reg))
                  )
                ) and (
                  (
                    (getsupreg(taicpu(hp2).oper[1]^.reg) = getsupreg(taicpu(hp1).oper[1]^.reg))
                  ) or (
                    (getsupreg(taicpu(hp2).oper[1]^.reg) = getsupreg(taicpu(p).oper[1]^.reg))
                  )
                )
              )
            )
          ) and (
            GetNextInstruction(hp2, hp3) and
            MatchInstruction(hp3, A_SHR,[]) and
            (taicpu(hp3).opsize = S_Q) and
            (taicpu(hp3).oper[0]^.typ = top_const) and (taicpu(hp2).oper[1]^.typ = top_reg) and
            (taicpu(hp3).oper[0]^.val = 1) and
            (taicpu(hp3).oper[1]^.reg = taicpu(hp2).oper[1]^.reg)
          ) then
          begin
            { Change   movl    x,    reg1d         movl    x,    reg1d
                       movl    y,    reg2d         movl    y,    reg2d
                       addq    reg2q,reg1q   or    leaq    (reg1q,reg2q),reg1q
                       shrq    $1,   reg1q         shrq    $1,   reg1q

            ( reg1d and reg2d can be switched around in the first two instructions )

              To       movl    x,    reg1d
                       addl    y,    reg1d
                       rcrl    $1,   reg1d

              This corresponds to the common expression (x + y) shr 1, where
              x and y are Cardinals (replacing "shr 1" with "div 2" produces
              smaller code, but won't account for x + y causing an overflow). [Kit]
            }

            if (getsupreg(taicpu(hp2).oper[1]^.reg) = getsupreg(taicpu(hp1).oper[1]^.reg)) then
              { Change first MOV command to have the same register as the final output }
              taicpu(p).oper[1]^.reg := taicpu(hp1).oper[1]^.reg
            else
              taicpu(hp1).oper[1]^.reg := taicpu(p).oper[1]^.reg;

            { Change second MOV command to an ADD command. This is easier than
              converting the existing command because it means we don't have to
              touch 'y', which might be a complicated reference, and also the
              fact that the third command might either be ADD or LEA. [Kit] }
            taicpu(hp1).opcode := A_ADD;

            { Delete old ADD/LEA instruction }
            RemoveInstruction(hp2);

            { Convert "shrq $1, reg1q" to "rcr $1, reg1d" }
            taicpu(hp3).opcode := A_RCR;
            taicpu(hp3).changeopsize(S_L);
            setsubreg(taicpu(hp3).oper[1]^.reg, R_SUBD);
{$endif x86_64}
          end;
      end;


    function TX86AsmOptimizer.OptPass2Imul(var p : tai) : boolean;
      var
        hp1 : tai;
      begin
        Result:=false;
        if (taicpu(p).ops >= 2) and
           ((taicpu(p).oper[0]^.typ = top_const) or
            ((taicpu(p).oper[0]^.typ = top_ref) and (taicpu(p).oper[0]^.ref^.refaddr=addr_full))) and
           (taicpu(p).oper[1]^.typ = top_reg) and
           ((taicpu(p).ops = 2) or
            ((taicpu(p).oper[2]^.typ = top_reg) and
             (taicpu(p).oper[2]^.reg = taicpu(p).oper[1]^.reg))) and
           GetLastInstruction(p,hp1) and
           MatchInstruction(hp1,A_MOV,[]) and
           MatchOpType(taicpu(hp1),top_reg,top_reg) and
           (taicpu(hp1).oper[1]^.reg = taicpu(p).oper[1]^.reg) then
          begin
            TransferUsedRegs(TmpUsedRegs);
            if not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,p,TmpUsedRegs)) or
              ((taicpu(p).ops = 3) and (taicpu(p).oper[1]^.reg=taicpu(p).oper[2]^.reg)) then
              { change
                  mov reg1,reg2
                  imul y,reg2 to imul y,reg1,reg2 }
              begin
                taicpu(p).ops := 3;
                taicpu(p).loadreg(2,taicpu(p).oper[1]^.reg);
                taicpu(p).loadreg(1,taicpu(hp1).oper[0]^.reg);
                DebugMsg(SPeepholeOptimization + 'MovImul2Imul done',p);
                RemoveInstruction(hp1);
                result:=true;
              end;
          end;
      end;


    procedure TX86AsmOptimizer.ConvertJumpToRET(const p: tai; const ret_p: tai);
      var
        ThisLabel: TAsmLabel;
      begin
        ThisLabel := tasmlabel(taicpu(p).oper[0]^.ref^.symbol);
        ThisLabel.decrefs;
        taicpu(p).opcode := A_RET;
        taicpu(p).is_jmp := false;
        taicpu(p).ops := taicpu(ret_p).ops;
        case taicpu(ret_p).ops of
          0:
            taicpu(p).clearop(0);
          1:
            taicpu(p).loadconst(0,taicpu(ret_p).oper[0]^.val);
          else
            internalerror(2016041301);
        end;

        { If the original label is now dead, it might turn out that the label
          immediately follows p.  As a result, everything beyond it, which will
          be just some final register configuration and a RET instruction, is
          now dead code. [Kit] }

        { NOTE: This is much faster than introducing a OptPass2RET routine and
          running RemoveDeadCodeAfterJump for each RET instruction, because
          this optimisation rarely happens and most RETs appear at the end of
          routines where there is nothing that can be stripped. [Kit] }
        if not ThisLabel.is_used then
          RemoveDeadCodeAfterJump(p);
      end;


    function TX86AsmOptimizer.OptPass2Jmp(var p : tai) : boolean;
      var
        hp1, hp2, hp3: tai;
        OperIdx: Integer;
      begin
        result:=false;
        if (taicpu(p).oper[0]^.typ=top_ref) and (taicpu(p).oper[0]^.ref^.refaddr=addr_full) and (taicpu(p).oper[0]^.ref^.base=NR_NO) and
          (taicpu(p).oper[0]^.ref^.index=NR_NO) and (taicpu(p).oper[0]^.ref^.symbol is tasmlabel) then
          begin
            hp1:=getlabelwithsym(tasmlabel(taicpu(p).oper[0]^.ref^.symbol));
            if (taicpu(p).condition=C_None) and assigned(hp1) and SkipLabels(hp1,hp1) and (hp1.typ = ait_instruction) then
              begin
                case taicpu(hp1).opcode of
                  A_RET:
                    {
                      change
                             jmp .L1
                             ...
                         .L1:
                             ret
                      into
                             ret
                    }
                    begin
                      ConvertJumpToRET(p, hp1);
                      result:=true;
                    end;
                  A_MOV:
                    {
                      change
                             jmp .L1
                             ...
                         .L1:
                             mov ##, ##
                             ret
                      into
                             mov ##, ##
                             ret
                    }
                    { This optimisation tends to increase code size if the pass 1 MOV optimisations aren't
                      re-run, so only do this particular optimisation if optimising for speed or when
                      optimisations are very in-depth. [Kit] }
                    if (current_settings.optimizerswitches * [cs_opt_level3, cs_opt_size]) <> [cs_opt_size] then
                      begin
                        GetNextInstruction(hp1, hp2);
                        if not Assigned(hp2) then
                          Exit;

                        if (hp2.typ in [ait_label, ait_align]) then
                          SkipLabels(hp2,hp2);
                        if Assigned(hp2) and MatchInstruction(hp2, A_RET, [S_NO]) then
                          begin
                            { Duplicate the MOV instruction }
                            hp3:=tai(hp1.getcopy);
                            asml.InsertBefore(hp3, p);

                            { Make sure the compiler knows about any final registers written here }
                            for OperIdx := 0 to 1 do
                              with taicpu(hp3).oper[OperIdx]^ do
                                begin
                                  case typ of
                                    top_ref:
                                      begin
                                        if (ref^.base <> NR_NO) {$ifdef x86_64} and (ref^.base <> NR_RIP) {$endif x86_64} then
                                          AllocRegBetween(ref^.base, hp3, tai(p.Next), UsedRegs);
                                        if (ref^.index <> NR_NO) {$ifdef x86_64} and (ref^.index <> NR_RIP) {$endif x86_64} then
                                          AllocRegBetween(ref^.index, hp3, tai(p.Next), UsedRegs);
                                      end;
                                    top_reg:
                                      AllocRegBetween(reg, hp3, tai(p.Next), UsedRegs);
                                    else
                                      ;
                                  end;
                                end;

                            { Now change the jump into a RET instruction }
                            ConvertJumpToRET(p, hp2);
                            result:=true;
                          end;
                      end;
                  else
                    ;
                end;
              end;
          end;
      end;


    class function TX86AsmOptimizer.CanBeCMOV(p : tai) : boolean;
      begin
         CanBeCMOV:=assigned(p) and
           MatchInstruction(p,A_MOV,[S_W,S_L,S_Q]) and
           { we can't use cmov ref,reg because
             ref could be nil and cmov still throws an exception
             if ref=nil but the mov isn't done (FK)
            or ((taicpu(p).oper[0]^.typ = top_ref) and
             (taicpu(p).oper[0]^.ref^.refaddr = addr_no))
           }
           (taicpu(p).oper[1]^.typ = top_reg) and
           (
             (taicpu(p).oper[0]^.typ = top_reg) or
             { allow references, but only pure symbols or got rel. addressing with RIP as based,
               it is not expected that this can cause a seg. violation }
             (
               (taicpu(p).oper[0]^.typ = top_ref) and
               IsRefSafe(taicpu(p).oper[0]^.ref)
             )
           );
      end;


    function TX86AsmOptimizer.OptPass2Jcc(var p : tai) : boolean;
      var
        hp1,hp2,hp3,hp4,hpmov2: tai;
        carryadd_opcode : TAsmOp;
        l : Longint;
        condition : TAsmCond;
        symbol: TAsmSymbol;
        reg: tsuperregister;
        regavailable: Boolean;
      begin
        result:=false;
        symbol:=nil;
        if GetNextInstruction(p,hp1) then
          begin
            symbol := TAsmLabel(taicpu(p).oper[0]^.ref^.symbol);

            if (hp1.typ=ait_instruction) and
               GetNextInstruction(hp1,hp2) and
               ((hp2.typ=ait_label) or
                 { trick to skip align }
                ((hp2.typ=ait_align) and GetNextInstruction(hp2,hp2) and (hp2.typ=ait_label))
               ) and
               (Tasmlabel(symbol) = Tai_label(hp2).labsym) then
                 { jb @@1                            cmc
                   inc/dec operand           -->     adc/sbb operand,0
                   @@1:

                   ... and ...

                   jnb @@1
                   inc/dec operand           -->     adc/sbb operand,0
                   @@1: }
              begin
                carryadd_opcode:=A_NONE;
                if Taicpu(p).condition in [C_NAE,C_B,C_C] then
                  begin
                    if (Taicpu(hp1).opcode=A_INC) or
                      ((Taicpu(hp1).opcode=A_ADD) and
                       MatchOptype(Taicpu(hp1),top_const,top_reg) and
                       (Taicpu(hp1).oper[0]^.val=1)
                      ) then
                      carryadd_opcode:=A_ADC;
                    if (Taicpu(hp1).opcode=A_DEC) or
                      ((Taicpu(hp1).opcode=A_SUB) and
                       MatchOptype(Taicpu(hp1),top_const,top_reg) and
                       (Taicpu(hp1).oper[0]^.val=1)
                      ) then
                      carryadd_opcode:=A_SBB;
                    if carryadd_opcode<>A_NONE then
                      begin
                        Taicpu(p).clearop(0);
                        Taicpu(p).ops:=0;
                        Taicpu(p).is_jmp:=false;
                        Taicpu(p).opcode:=A_CMC;
                        Taicpu(p).condition:=C_NONE;
                        DebugMsg(SPeepholeOptimization+'JccAdd/Inc/Dec2CmcAdc/Sbb',p);
                        Taicpu(hp1).ops:=2;
                        if (Taicpu(hp1).opcode=A_ADD) or (Taicpu(hp1).opcode=A_SUB) then
                          Taicpu(hp1).loadoper(1,Taicpu(hp1).oper[1]^)
                        else
                          Taicpu(hp1).loadoper(1,Taicpu(hp1).oper[0]^);
                        Taicpu(hp1).loadconst(0,0);
                        Taicpu(hp1).opcode:=carryadd_opcode;
                        result:=true;
                        exit;
                      end;
                  end
                else if Taicpu(p).condition in [C_AE,C_NB,C_NC] then
                  begin
                    if (Taicpu(hp1).opcode=A_INC) or
                      ((Taicpu(hp1).opcode=A_ADD) and
                       MatchOptype(Taicpu(hp1),top_const,top_reg) and
                       (Taicpu(hp1).oper[0]^.val=1)
                      ) then
                      carryadd_opcode:=A_ADC;
                    if (Taicpu(hp1).opcode=A_DEC) or
                      ((Taicpu(hp1).opcode=A_SUB) and
                       MatchOptype(Taicpu(hp1),top_const,top_reg) and
                       (Taicpu(hp1).oper[0]^.val=1)
                      ) then
                      carryadd_opcode:=A_SBB;
                    if carryadd_opcode<>A_NONE then
                      begin
                        Taicpu(hp1).ops:=2;
                        DebugMsg(SPeepholeOptimization+'JccAdd/Inc/Dec2Adc/Sbb',p);
                        if (Taicpu(hp1).opcode=A_ADD) or (Taicpu(hp1).opcode=A_SUB) then
                          Taicpu(hp1).loadoper(1,Taicpu(hp1).oper[1]^)
                        else
                          Taicpu(hp1).loadoper(1,Taicpu(hp1).oper[0]^);
                        Taicpu(hp1).loadconst(0,0);
                        Taicpu(hp1).opcode:=carryadd_opcode;
                        RemoveCurrentP(p, hp1);
                        result:=true;
                        exit;
                      end;
                  end
                 {
                   jcc @@1                            setcc tmpreg
                   inc/dec/add/sub operand    ->      (movzx tmpreg)
                   @@1:                               add/sub tmpreg,operand

                   While this increases code size slightly, it makes the code much faster if the
                   jump is unpredictable
                 }
                else if not(cs_opt_size in current_settings.optimizerswitches) and
                  ((((Taicpu(hp1).opcode=A_ADD) or (Taicpu(hp1).opcode=A_SUB)) and
                    (Taicpu(hp1).oper[0]^.typ=top_const) and
                    (Taicpu(hp1).oper[1]^.typ=top_reg) and
                    (Taicpu(hp1).oper[0]^.val=1)) or
                   ((Taicpu(hp1).opcode=A_INC) or (Taicpu(hp1).opcode=A_DEC))
                  ) then
                  begin
                    TransferUsedRegs(TmpUsedRegs);
                    UpdateUsedRegs(TmpUsedRegs, tai(p.next));

                    { search for an available register which is volatile }
                    regavailable:=false;
                    for reg in tcpuregisterset do
                      begin
                        if (reg in paramanager.get_volatile_registers_int(current_procinfo.procdef.proccalloption)) and
                          not(reg in TmpUsedRegs[R_INTREGISTER].GetUsedRegs) and
                          not(RegInInstruction(newreg(R_INTREGISTER,reg,R_SUBL),hp1))
{$ifdef i386}
                          and (reg in [RS_EAX,RS_EBX,RS_ECX,RS_EDX])
{$endif i386}
                          then
                          begin
                            regavailable:=true;
                            break;
                          end;
                      end;

                    if regavailable then
                      begin
                        Taicpu(p).clearop(0);
                        Taicpu(p).ops:=1;
                        Taicpu(p).is_jmp:=false;
                        Taicpu(p).opcode:=A_SETcc;
                        DebugMsg(SPeepholeOptimization+'JccAdd2SetccAdd',p);
                        Taicpu(p).condition:=inverse_cond(Taicpu(p).condition);
                        Taicpu(p).loadreg(0,newreg(R_INTREGISTER,reg,R_SUBL));

                        if getsubreg(Taicpu(hp1).oper[1]^.reg)<>R_SUBL then
                          begin
                            case getsubreg(Taicpu(hp1).oper[1]^.reg) of
                              R_SUBW:
                                hp2:=Taicpu.op_reg_reg(A_MOVZX,S_BW,newreg(R_INTREGISTER,reg,R_SUBL),
                                  newreg(R_INTREGISTER,reg,R_SUBW));
                              R_SUBD,
                              R_SUBQ:
                                hp2:=Taicpu.op_reg_reg(A_MOVZX,S_BL,newreg(R_INTREGISTER,reg,R_SUBL),
                                  newreg(R_INTREGISTER,reg,R_SUBD));
                              else
                                Internalerror(2020030601);
                            end;
                            taicpu(hp2).fileinfo:=taicpu(hp1).fileinfo;
                            asml.InsertAfter(hp2,p);
                          end;
                        if (Taicpu(hp1).opcode=A_INC) or (Taicpu(hp1).opcode=A_DEC) then
                          begin
                            Taicpu(hp1).ops:=2;
                            Taicpu(hp1).loadoper(1,Taicpu(hp1).oper[0]^)
                          end;
                        Taicpu(hp1).loadreg(0,newreg(R_INTREGISTER,reg,getsubreg(Taicpu(hp1).oper[1]^.reg)));
                        AllocRegBetween(newreg(R_INTREGISTER,reg,getsubreg(Taicpu(hp1).oper[1]^.reg)),p,hp1,UsedRegs);
                      end;
                  end;
              end;

          { Detect the following:
              jmp<cond>     @Lbl1
              jmp           @Lbl2
              ...
            @Lbl1:
              ret

            Change to:

              jmp<inv_cond> @Lbl2
              ret
          }
            if MatchInstruction(hp1,A_JMP,[]) and (taicpu(hp1).oper[0]^.ref^.refaddr=addr_full) then
              begin
                hp2:=getlabelwithsym(TAsmLabel(symbol));
                if Assigned(hp2) and SkipLabels(hp2,hp2) and
                  MatchInstruction(hp2,A_RET,[S_NO]) then
                  begin
                    taicpu(p).condition := inverse_cond(taicpu(p).condition);

                    { Change label address to that of the unconditional jump }
                    taicpu(p).loadoper(0, taicpu(hp1).oper[0]^);

                    TAsmLabel(symbol).DecRefs;
                    taicpu(hp1).opcode := A_RET;
                    taicpu(hp1).is_jmp := false;
                    taicpu(hp1).ops := taicpu(hp2).ops;
                    DebugMsg(SPeepholeOptimization+'JccJmpRet2J!ccRet',p);
                    case taicpu(hp2).ops of
                      0:
                        taicpu(hp1).clearop(0);
                      1:
                        taicpu(hp1).loadconst(0,taicpu(hp2).oper[0]^.val);
                      else
                        internalerror(2016041302);
                    end;
                  end;
              end;
          end;
{$ifndef i8086}
        if CPUX86_HAS_CMOV in cpu_capabilities[current_settings.cputype] then
          begin
             { check for
                    jCC   xxx
                    <several movs>
                 xxx:
             }
             l:=0;
             GetNextInstruction(p, hp1);
             while assigned(hp1) and
               CanBeCMOV(hp1) and
               { stop on labels }
               not(hp1.typ=ait_label) do
               begin
                  inc(l);
                  GetNextInstruction(hp1,hp1);
               end;
             if assigned(hp1) then
               begin
                  if FindLabel(tasmlabel(symbol),hp1) then
                    begin
                      if (l<=4) and (l>0) then
                        begin
                          condition:=inverse_cond(taicpu(p).condition);
                          GetNextInstruction(p,hp1);
                          repeat
                            if not Assigned(hp1) then
                              InternalError(2018062900);

                            taicpu(hp1).opcode:=A_CMOVcc;
                            taicpu(hp1).condition:=condition;
                            UpdateUsedRegs(hp1);
                            GetNextInstruction(hp1,hp1);
                          until not(CanBeCMOV(hp1));

                          { Remember what hp1 is in case there's multiple aligns to get rid of }
                          hp2 := hp1;
                          repeat
                            if not Assigned(hp2) then
                              InternalError(2018062910);

                            case hp2.typ of
                              ait_label:
                                { What we expected - break out of the loop (it won't be a dead label at the top of
                                  a cluster because that was optimised at an earlier stage) }
                                Break;
                              ait_align:
                                { Go to the next entry until a label is found (may be multiple aligns before it) }
                                begin
                                  hp2 := tai(hp2.Next);
                                  Continue;
                                end;
                              else
                                begin
                                  { Might be a comment or temporary allocation entry }
                                  if not (hp2.typ in SkipInstr) then
                                    InternalError(2018062911);

                                  hp2 := tai(hp2.Next);
                                  Continue;
                                end;
                            end;

                          until False;

                          { Now we can safely decrement the reference count }
                          tasmlabel(symbol).decrefs;

                          DebugMsg(SPeepholeOptimization+'JccMov2CMov',p);

                          { Remove the original jump }
                          RemoveInstruction(p); { Note, the choice to not use RemoveCurrentp is deliberate }

                          GetNextInstruction(hp2, p); { Instruction after the label }

                          { Remove the label if this is its final reference }
                          if (tasmlabel(symbol).getrefs=0) then
                            StripLabelFast(hp1);

                          if Assigned(p) then
                            begin
                              UpdateUsedRegs(p);
                              result:=true;
                            end;
                          exit;
                        end;
                    end
                  else
                    begin
                       { check further for
                              jCC   xxx
                              <several movs 1>
                              jmp   yyy
                      xxx:
                              <several movs 2>
                      yyy:
                       }
                      { hp2 points to jmp yyy }
                      hp2:=hp1;
                      { skip hp1 to xxx (or an align right before it) }
                      GetNextInstruction(hp1, hp1);

                      if assigned(hp2) and
                        assigned(hp1) and
                        (l<=3) and
                        (hp2.typ=ait_instruction) and
                        (taicpu(hp2).is_jmp) and
                        (taicpu(hp2).condition=C_None) and
                        { real label and jump, no further references to the
                          label are allowed }
                        (tasmlabel(symbol).getrefs=1) and
                        FindLabel(tasmlabel(symbol),hp1) then
                         begin
                           l:=0;
                           { skip hp1 to <several moves 2> }
                           if (hp1.typ = ait_align) then
                             GetNextInstruction(hp1, hp1);

                           GetNextInstruction(hp1, hpmov2);

                           hp1 := hpmov2;
                           while assigned(hp1) and
                             CanBeCMOV(hp1) do
                             begin
                               inc(l);
                               GetNextInstruction(hp1, hp1);
                             end;
                           { hp1 points to yyy (or an align right before it) }
                           hp3 := hp1;
                           if assigned(hp1) and
                             FindLabel(tasmlabel(taicpu(hp2).oper[0]^.ref^.symbol),hp1) then
                             begin
                                condition:=inverse_cond(taicpu(p).condition);
                                GetNextInstruction(p,hp1);
                                repeat
                                  taicpu(hp1).opcode:=A_CMOVcc;
                                  taicpu(hp1).condition:=condition;
                                  UpdateUsedRegs(hp1);
                                  GetNextInstruction(hp1,hp1);
                                until not(assigned(hp1)) or
                                  not(CanBeCMOV(hp1));

                                condition:=inverse_cond(condition);
                                hp1 := hpmov2;
                                { hp1 is now at <several movs 2> }
                                while Assigned(hp1) and CanBeCMOV(hp1) do
                                  begin
                                    taicpu(hp1).opcode:=A_CMOVcc;
                                    taicpu(hp1).condition:=condition;
                                    UpdateUsedRegs(hp1);
                                    GetNextInstruction(hp1,hp1);
                                  end;

                                hp1 := p;

                                { Get first instruction after label }
                                GetNextInstruction(hp3, p);

                                if assigned(p) and (hp3.typ = ait_align) then
                                  GetNextInstruction(p, p);

                                { Don't dereference yet, as doing so will cause
                                  GetNextInstruction to skip the label and
                                  optional align marker. [Kit] }
                                GetNextInstruction(hp2, hp4);

                                DebugMsg(SPeepholeOptimization+'JccMovJmpMov2CMovCMov',hp1);

                                { remove jCC }
                                RemoveInstruction(hp1);

                                { Now we can safely decrement it }
                                tasmlabel(symbol).decrefs;

                                { Remove label xxx (it will have a ref of zero due to the initial check }
                                StripLabelFast(hp4);

                                { remove jmp }
                                symbol := taicpu(hp2).oper[0]^.ref^.symbol;

                                RemoveInstruction(hp2);

                                { As before, now we can safely decrement it }
                                tasmlabel(symbol).decrefs;

                                { Remove label yyy (and the optional alignment) if its reference falls to zero }
                                if tasmlabel(symbol).getrefs = 0 then
                                  StripLabelFast(hp3);

                                if Assigned(p) then
                                  begin
                                    UpdateUsedRegs(p);
                                    result:=true;
                                  end;
                                exit;
                             end;
                         end;
                    end;
               end;
          end;
{$endif i8086}
      end;


    function TX86AsmOptimizer.OptPass1Movx(var p : tai) : boolean;
      var
        hp1,hp2: tai;
        reg_and_hp1_is_instr: Boolean;
      begin
        result:=false;
        reg_and_hp1_is_instr:=(taicpu(p).oper[1]^.typ = top_reg) and
          GetNextInstruction(p,hp1) and
          (hp1.typ = ait_instruction);

        if reg_and_hp1_is_instr and
          (
            (taicpu(hp1).opcode <> A_LEA) or
            { If the LEA instruction can be converted into an arithmetic instruction,
              it may be possible to then fold it. }
            (
              { If the flags register is in use, don't change the instruction
                to an ADD otherwise this will scramble the flags. [Kit] }
              not RegInUsedRegs(NR_DEFAULTFLAGS, UsedRegs) and
              ConvertLEA(taicpu(hp1))
            )
          ) and
          IsFoldableArithOp(taicpu(hp1),taicpu(p).oper[1]^.reg) and
          GetNextInstruction(hp1,hp2) and
          MatchInstruction(hp2,A_MOV,[]) and
          (taicpu(hp2).oper[0]^.typ = top_reg) and
          OpsEqual(taicpu(hp2).oper[1]^,taicpu(p).oper[0]^) and
          ((taicpu(p).opsize in [S_BW,S_BL]) and (taicpu(hp2).opsize=S_B) or
           (taicpu(p).opsize in [S_WL]) and (taicpu(hp2).opsize=S_W)) and
{$ifdef i386}
           { not all registers have byte size sub registers on i386 }
           ((taicpu(hp2).opsize<>S_B) or (getsupreg(taicpu(hp1).oper[0]^.reg) in [RS_EAX, RS_EBX, RS_ECX, RS_EDX])) and
{$endif i386}
           (((taicpu(hp1).ops=2) and
             (getsupreg(taicpu(hp2).oper[0]^.reg)=getsupreg(taicpu(hp1).oper[1]^.reg))) or
             ((taicpu(hp1).ops=1) and
             (getsupreg(taicpu(hp2).oper[0]^.reg)=getsupreg(taicpu(hp1).oper[0]^.reg)))) and
           not(RegUsedAfterInstruction(taicpu(hp2).oper[0]^.reg,hp2,UsedRegs)) then
          begin
            { change   movsX/movzX    reg/ref, reg2
                       add/sub/or/... reg3/$const, reg2
                       mov            reg2 reg/ref
              to       add/sub/or/... reg3/$const, reg/ref      }

            { by example:
                movswl  %si,%eax        movswl  %si,%eax      p
                decl    %eax            addl    %edx,%eax     hp1
                movw    %ax,%si         movw    %ax,%si       hp2
              ->
                movswl  %si,%eax        movswl  %si,%eax      p
                decw    %eax            addw    %edx,%eax     hp1
                movw    %ax,%si         movw    %ax,%si       hp2
            }
            taicpu(hp1).changeopsize(taicpu(hp2).opsize);
            {
              ->
                movswl  %si,%eax        movswl  %si,%eax      p
                decw    %si             addw    %dx,%si       hp1
                movw    %ax,%si         movw    %ax,%si       hp2
            }
            case taicpu(hp1).ops of
              1:
               taicpu(hp1).loadoper(0,taicpu(hp2).oper[1]^);
              2:
                begin
                  taicpu(hp1).loadoper(1,taicpu(hp2).oper[1]^);
                  if (taicpu(hp1).oper[0]^.typ = top_reg) then
                    setsubreg(taicpu(hp1).oper[0]^.reg,getsubreg(taicpu(hp2).oper[0]^.reg));
                end;
              else
                internalerror(2008042701);
            end;
            {
              ->
                decw    %si             addw    %dx,%si       p
            }
            DebugMsg(SPeepholeOptimization + 'var3',p);
            RemoveCurrentP(p, hp1);
            RemoveInstruction(hp2);
          end
        else if reg_and_hp1_is_instr and
          (taicpu(hp1).opcode = A_MOV) and
          MatchOpType(taicpu(hp1),top_reg,top_reg) and
          (MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[0]^)
{$ifdef x86_64}
           { check for implicit extension to 64 bit }
           or
           ((taicpu(p).opsize in [S_BL,S_WL]) and
            (taicpu(hp1).opsize=S_Q) and
            SuperRegistersEqual(taicpu(p).oper[1]^.reg,taicpu(hp1).oper[0]^.reg)
           )
{$endif x86_64}
          )
          then
          begin
            { change
              movx   %reg1,%reg2
              mov    %reg2,%reg3
              dealloc %reg2

              into

              movx   %reg,%reg3
            }
            TransferUsedRegs(TmpUsedRegs);
            UpdateUsedRegs(TmpUsedRegs, tai(p.next));
            if not(RegUsedAfterInstruction(taicpu(p).oper[1]^.reg,hp1,TmpUsedRegs)) then
              begin
                DebugMsg(SPeepholeOptimization + 'MovxMov2Movx',p);
{$ifdef x86_64}
                if (taicpu(p).opsize in [S_BL,S_WL]) and
                  (taicpu(hp1).opsize=S_Q) then
                  taicpu(p).loadreg(1,newreg(R_INTREGISTER,getsupreg(taicpu(hp1).oper[1]^.reg),R_SUBD))
                else
{$endif x86_64}
                  taicpu(p).loadreg(1,taicpu(hp1).oper[1]^.reg);
                RemoveInstruction(hp1);
              end;
          end
        else if reg_and_hp1_is_instr and
          (taicpu(p).oper[0]^.typ = top_reg) and
          (
            (taicpu(hp1).opcode = A_SHL) or (taicpu(hp1).opcode = A_SAL)
          ) and
          (taicpu(hp1).oper[0]^.typ = top_const) and
          SuperRegistersEqual(taicpu(p).oper[0]^.reg, taicpu(p).oper[1]^.reg) and
          MatchOperand(taicpu(hp1).oper[1]^, taicpu(p).oper[1]^.reg) and
          { Minimum shift value allowed is the bit difference between the sizes }
          (taicpu(hp1).oper[0]^.val >=
            { Multiply by 8 because tcgsize2size returns bytes, not bits }
            8 * (
              tcgsize2size[reg_cgsize(taicpu(p).oper[1]^.reg)] -
              tcgsize2size[reg_cgsize(taicpu(p).oper[0]^.reg)]
            )
          ) then
          begin
            { For:
                movsx/movzx %reg1,%reg1 (same register, just different sizes)
                shl/sal     ##,   %reg1

              Remove the movsx/movzx instruction if the shift overwrites the
              extended bits of the register (e.g. movslq %eax,%rax; shlq $32,%rax
            }
            DebugMsg(SPeepholeOptimization + 'MovxShl2Shl',p);
            RemoveCurrentP(p, hp1);
            Result := True;
            Exit;
          end
        else if taicpu(p).opcode=A_MOVZX then
          begin
            { removes superfluous And's after movzx's }
            if reg_and_hp1_is_instr and
              (taicpu(hp1).opcode = A_AND) and
              MatchOpType(taicpu(hp1),top_const,top_reg) and
              ((taicpu(hp1).oper[1]^.reg = taicpu(p).oper[1]^.reg)
{$ifdef x86_64}
               { check for implicit extension to 64 bit }
               or
               ((taicpu(p).opsize in [S_BL,S_WL]) and
                (taicpu(hp1).opsize=S_Q) and
                SuperRegistersEqual(taicpu(p).oper[1]^.reg,taicpu(hp1).oper[1]^.reg)
               )
{$endif x86_64}
              )
              then
              begin
                case taicpu(p).opsize Of
                  S_BL, S_BW{$ifdef x86_64}, S_BQ{$endif x86_64}:
                    if (taicpu(hp1).oper[0]^.val = $ff) then
                      begin
                        DebugMsg(SPeepholeOptimization + 'MovzAnd2Movz1',p);
                        RemoveInstruction(hp1);
                        Result:=true;
                        exit;
                      end;
                    S_WL{$ifdef x86_64}, S_WQ{$endif x86_64}:
                      if (taicpu(hp1).oper[0]^.val = $ffff) then
                        begin
                          DebugMsg(SPeepholeOptimization + 'MovzAnd2Movz2',p);
                          RemoveInstruction(hp1);
                          Result:=true;
                          exit;
                        end;
{$ifdef x86_64}
                    S_LQ:
                      if (taicpu(hp1).oper[0]^.val = $ffffffff) then
                        begin
                          DebugMsg(SPeepholeOptimization + 'MovzAnd2Movz3',p);
                          RemoveInstruction(hp1);
                          Result:=true;
                          exit;
                        end;
{$endif x86_64}
                    else
                      ;
                end;
                { we cannot get rid of the and, but can we get rid of the movz ?}
                if SuperRegistersEqual(taicpu(p).oper[0]^.reg,taicpu(p).oper[1]^.reg) then
                  begin
                    case taicpu(p).opsize Of
                      S_BL, S_BW{$ifdef x86_64}, S_BQ{$endif x86_64}:
                        if (taicpu(hp1).oper[0]^.val and $ff)=taicpu(hp1).oper[0]^.val then
                          begin
                            DebugMsg(SPeepholeOptimization + 'MovzAnd2And1',p);
                            RemoveCurrentP(p,hp1);
                            Result:=true;
                            exit;
                          end;
                        S_WL{$ifdef x86_64}, S_WQ{$endif x86_64}:
                          if (taicpu(hp1).oper[0]^.val and $ffff)=taicpu(hp1).oper[0]^.val then
                            begin
                              DebugMsg(SPeepholeOptimization + 'MovzAnd2And2',p);
                              RemoveCurrentP(p,hp1);
                              Result:=true;
                              exit;
                            end;
{$ifdef x86_64}
                        S_LQ:
                          if (taicpu(hp1).oper[0]^.val and $ffffffff)=taicpu(hp1).oper[0]^.val then
                            begin
                              DebugMsg(SPeepholeOptimization + 'MovzAnd2And3',p);
                              RemoveCurrentP(p,hp1);
                              Result:=true;
                              exit;
                            end;
{$endif x86_64}
                        else
                          ;
                    end;
                end;
              end;
            { changes some movzx constructs to faster synonyms (all examples
              are given with eax/ax, but are also valid for other registers)}
            if MatchOpType(taicpu(p),top_reg,top_reg) then
                begin
                  case taicpu(p).opsize of
                    { Technically, movzbw %al,%ax cannot be encoded in 32/64-bit mode
                      (the machine code is equivalent to movzbl %al,%eax), but the
                      code generator still generates that assembler instruction and
                      it is silently converted.  This should probably be checked.
                      [Kit] }
                    S_BW:
                      begin
                        if (getsupreg(taicpu(p).oper[0]^.reg)=getsupreg(taicpu(p).oper[1]^.reg)) and
                          (
                            not IsMOVZXAcceptable
                            { and $0xff,%ax has a smaller encoding but risks a partial write penalty }
                            or (
                              (cs_opt_size in current_settings.optimizerswitches) and
                              (taicpu(p).oper[1]^.reg = NR_AX)
                            )
                          ) then
                          {Change "movzbw %al, %ax" to "andw $0x0ffh, %ax"}
                          begin
                            DebugMsg(SPeepholeOptimization + 'var7',p);
                            taicpu(p).opcode := A_AND;
                            taicpu(p).changeopsize(S_W);
                            taicpu(p).loadConst(0,$ff);
                            Result := True;
                          end
                        else if not IsMOVZXAcceptable and
                          GetNextInstruction(p, hp1) and
                          (tai(hp1).typ = ait_instruction) and
                          (taicpu(hp1).opcode = A_AND) and
                          MatchOpType(taicpu(hp1),top_const,top_reg) and
                          (taicpu(hp1).oper[1]^.reg = taicpu(p).oper[1]^.reg) then
                        { Change "movzbw %reg1, %reg2; andw $const, %reg2"
                          to "movw %reg1, reg2; andw $(const1 and $ff), %reg2"}
                          begin
                            DebugMsg(SPeepholeOptimization + 'var8',p);
                            taicpu(p).opcode := A_MOV;
                            taicpu(p).changeopsize(S_W);
                            setsubreg(taicpu(p).oper[0]^.reg,R_SUBW);
                            taicpu(hp1).loadConst(0,taicpu(hp1).oper[0]^.val and $ff);
                            Result := True;
                          end;
                      end;
{$ifndef i8086} { movzbl %al,%eax cannot be encoded in 16-bit mode (the machine code is equivalent to movzbw %al,%ax }
                    S_BL:
                      begin
                        if (getsupreg(taicpu(p).oper[0]^.reg)=getsupreg(taicpu(p).oper[1]^.reg)) and
                          (
                            not IsMOVZXAcceptable
                            { and $0xff,%eax has a smaller encoding but risks a partial write penalty }
                            or (
                              (cs_opt_size in current_settings.optimizerswitches) and
                              (taicpu(p).oper[1]^.reg = NR_EAX)
                            )
                          ) then
                          { Change "movzbl %al, %eax" to "andl $0x0ffh, %eax" }
                          begin
                            DebugMsg(SPeepholeOptimization + 'var9',p);
                            taicpu(p).opcode := A_AND;
                            taicpu(p).changeopsize(S_L);
                            taicpu(p).loadConst(0,$ff);
                            Result := True;
                          end
                        else if not IsMOVZXAcceptable and
                          GetNextInstruction(p, hp1) and
                          (tai(hp1).typ = ait_instruction) and
                          (taicpu(hp1).opcode = A_AND) and
                          MatchOpType(taicpu(hp1),top_const,top_reg) and
                          (taicpu(hp1).oper[1]^.reg = taicpu(p).oper[1]^.reg) then
                          { Change "movzbl %reg1, %reg2; andl $const, %reg2"
                            to "movl %reg1, reg2; andl $(const1 and $ff), %reg2"}
                          begin
                            DebugMsg(SPeepholeOptimization + 'var10',p);
                            taicpu(p).opcode := A_MOV;
                            taicpu(p).changeopsize(S_L);
                            { do not use R_SUBWHOLE
                              as movl %rdx,%eax
                              is invalid in assembler PM }
                            setsubreg(taicpu(p).oper[0]^.reg, R_SUBD);
                            taicpu(hp1).loadConst(0,taicpu(hp1).oper[0]^.val and $ff);
                            Result := True;
                          end;
                      end;
{$endif i8086}
                    S_WL:
                      if not IsMOVZXAcceptable then
                        begin
                          if (getsupreg(taicpu(p).oper[0]^.reg)=getsupreg(taicpu(p).oper[1]^.reg)) then
                            { Change "movzwl %ax, %eax" to "andl $0x0ffffh, %eax" }
                            begin
                              DebugMsg(SPeepholeOptimization + 'var11',p);
                              taicpu(p).opcode := A_AND;
                              taicpu(p).changeopsize(S_L);
                              taicpu(p).loadConst(0,$ffff);
                              Result := True;
                            end
                          else if GetNextInstruction(p, hp1) and
                            (tai(hp1).typ = ait_instruction) and
                            (taicpu(hp1).opcode = A_AND) and
                            (taicpu(hp1).oper[0]^.typ = top_const) and
                            (taicpu(hp1).oper[1]^.typ = top_reg) and
                            (taicpu(hp1).oper[1]^.reg = taicpu(p).oper[1]^.reg) then
                            { Change "movzwl %reg1, %reg2; andl $const, %reg2"
                              to "movl %reg1, reg2; andl $(const1 and $ffff), %reg2"}
                            begin
                              DebugMsg(SPeepholeOptimization + 'var12',p);
                              taicpu(p).opcode := A_MOV;
                              taicpu(p).changeopsize(S_L);
                              { do not use R_SUBWHOLE
                                as movl %rdx,%eax
                                is invalid in assembler PM }
                              setsubreg(taicpu(p).oper[0]^.reg, R_SUBD);
                              taicpu(hp1).loadConst(0,taicpu(hp1).oper[0]^.val and $ffff);
                              Result := True;
                            end;
                        end;
                    else
                      InternalError(2017050705);
                  end;
                end
              else if not IsMOVZXAcceptable and (taicpu(p).oper[0]^.typ = top_ref) then
                  begin
                    if GetNextInstruction(p, hp1) and
                      (tai(hp1).typ = ait_instruction) and
                      (taicpu(hp1).opcode = A_AND) and
                      MatchOpType(taicpu(hp1),top_const,top_reg) and
                      (taicpu(hp1).oper[1]^.reg = taicpu(p).oper[1]^.reg) then
                      begin
                        //taicpu(p).opcode := A_MOV;
                        case taicpu(p).opsize Of
                          S_BL:
                            begin
                              DebugMsg(SPeepholeOptimization + 'var13',p);
                              taicpu(hp1).changeopsize(S_L);
                              taicpu(hp1).loadConst(0,taicpu(hp1).oper[0]^.val and $ff);
                            end;
                          S_WL:
                            begin
                              DebugMsg(SPeepholeOptimization + 'var14',p);
                              taicpu(hp1).changeopsize(S_L);
                              taicpu(hp1).loadConst(0,taicpu(hp1).oper[0]^.val and $ffff);
                            end;
                          S_BW:
                            begin
                              DebugMsg(SPeepholeOptimization + 'var15',p);
                              taicpu(hp1).changeopsize(S_W);
                              taicpu(hp1).loadConst(0,taicpu(hp1).oper[0]^.val and $ff);
                            end;
                          else
                            Internalerror(2017050704)
                        end;
                        Result := True;
                      end;
                  end;
          end;
      end;


    function TX86AsmOptimizer.OptPass1AND(var p : tai) : boolean;
      var
        hp1 : tai;
        MaskLength : Cardinal;
      begin
        Result:=false;

        if GetNextInstruction(p, hp1) then
          begin
            if MatchOpType(taicpu(p),top_const,top_reg) and
              MatchInstruction(hp1,A_AND,[]) and
              MatchOpType(taicpu(hp1),top_const,top_reg) and
              (getsupreg(taicpu(p).oper[1]^.reg) = getsupreg(taicpu(hp1).oper[1]^.reg)) and
              { the second register must contain the first one, so compare their subreg types }
              (getsubreg(taicpu(p).oper[1]^.reg)<=getsubreg(taicpu(hp1).oper[1]^.reg)) and
              (abs(taicpu(p).oper[0]^.val and taicpu(hp1).oper[0]^.val)<$80000000) then
              { change
                  and const1, reg
                  and const2, reg
                to
                  and (const1 and const2), reg
              }
              begin
                taicpu(hp1).loadConst(0, taicpu(p).oper[0]^.val and taicpu(hp1).oper[0]^.val);
                DebugMsg(SPeepholeOptimization + 'AndAnd2And done',hp1);
                RemoveCurrentP(p, hp1);
                Result:=true;
                exit;
              end
            else if MatchOpType(taicpu(p),top_const,top_reg) and
              MatchInstruction(hp1,A_MOVZX,[]) and
              MatchOpType(taicpu(hp1),top_reg,top_reg) and
              SuperRegistersEqual(taicpu(p).oper[1]^.reg,taicpu(hp1).oper[1]^.reg) and
              (getsupreg(taicpu(hp1).oper[0]^.reg)=getsupreg(taicpu(hp1).oper[1]^.reg)) and
               (((taicpu(p).opsize=S_W) and
                 (taicpu(hp1).opsize=S_BW)) or
                ((taicpu(p).opsize=S_L) and
                 (taicpu(hp1).opsize in [S_WL,S_BL{$ifdef x86_64},S_BQ,S_WQ{$endif x86_64}]))
{$ifdef x86_64}
                  or
                 ((taicpu(p).opsize=S_Q) and
                  (taicpu(hp1).opsize in [S_BQ,S_WQ,S_BL,S_WL]))
{$endif x86_64}
                ) then
                  begin
                    if (((taicpu(hp1).opsize) in [S_BW,S_BL{$ifdef x86_64},S_BQ{$endif x86_64}]) and
                        ((taicpu(p).oper[0]^.val and $ff)=taicpu(p).oper[0]^.val)
                         ) or
                       (((taicpu(hp1).opsize) in [S_WL{$ifdef x86_64},S_WQ{$endif x86_64}]) and
                        ((taicpu(p).oper[0]^.val and $ffff)=taicpu(p).oper[0]^.val))
                    then
                      begin
                        { Unlike MOVSX, MOVZX doesn't actually have a version that zero-extends a
                          32-bit register to a 64-bit register, or even a version called MOVZXD, so
                          code that tests for the presence of AND 0xffffffff followed by MOVZX is
                          wasted, and is indictive of a compiler bug if it were triggered. [Kit]

                          NOTE: To zero-extend from 32 bits to 64 bits, simply use the standard MOV.
                        }
                        DebugMsg(SPeepholeOptimization + 'AndMovzToAnd done',p);

                        RemoveInstruction(hp1);
                        Exit;
                      end;
                  end
            else if MatchOpType(taicpu(p),top_const,top_reg) and
              MatchInstruction(hp1,A_SHL,[]) and
              MatchOpType(taicpu(hp1),top_const,top_reg) and
              (getsupreg(taicpu(p).oper[1]^.reg)=getsupreg(taicpu(hp1).oper[1]^.reg)) then
              begin
{$ifopt R+}
{$define RANGE_WAS_ON}
{$R-}
{$endif}
                { get length of potential and mask }
                MaskLength:=SizeOf(taicpu(p).oper[0]^.val)*8-BsrQWord(taicpu(p).oper[0]^.val)-1;

                { really a mask? }
{$ifdef RANGE_WAS_ON}
{$R+}
{$endif}
                if (((QWord(1) shl MaskLength)-1)=taicpu(p).oper[0]^.val) and
                  { unmasked part shifted out? }
                  ((MaskLength+taicpu(hp1).oper[0]^.val)>=topsize2memsize[taicpu(hp1).opsize]) then
                  begin
                    DebugMsg(SPeepholeOptimization + 'AndShlToShl done',p);
                    RemoveCurrentP(p, hp1);
                    Result:=true;
                    exit;
                  end;
              end
            else if MatchOpType(taicpu(p),top_const,top_reg) and
              MatchInstruction(hp1,A_MOVSX{$ifdef x86_64},A_MOVSXD{$endif x86_64},[]) and
              (taicpu(hp1).oper[0]^.typ = top_reg) and
              MatchOperand(taicpu(p).oper[1]^,taicpu(hp1).oper[1]^) and
              (getsupreg(taicpu(hp1).oper[0]^.reg)=getsupreg(taicpu(hp1).oper[1]^.reg)) and
               (((taicpu(p).opsize=S_W) and
                 (taicpu(hp1).opsize=S_BW)) or
                ((taicpu(p).opsize=S_L) and
                 (taicpu(hp1).opsize in [S_WL,S_BL]))
{$ifdef x86_64}
                 or
                 ((taicpu(p).opsize=S_Q) and
                 (taicpu(hp1).opsize in [S_BQ,S_WQ,S_LQ]))
{$endif x86_64}
                ) then
                  begin
                    if (((taicpu(hp1).opsize) in [S_BW,S_BL{$ifdef x86_64},S_BQ{$endif x86_64}]) and
                        ((taicpu(p).oper[0]^.val and $7f)=taicpu(p).oper[0]^.val)
                         ) or
                       (((taicpu(hp1).opsize) in [S_WL{$ifdef x86_64},S_WQ{$endif x86_64}]) and
                        ((taicpu(p).oper[0]^.val and $7fff)=taicpu(p).oper[0]^.val))
{$ifdef x86_64}
                       or
                       (((taicpu(hp1).opsize)=S_LQ) and
                        ((taicpu(p).oper[0]^.val and $7fffffff)=taicpu(p).oper[0]^.val)
                       )
{$endif x86_64}
                       then
                       begin
                         DebugMsg(SPeepholeOptimization + 'AndMovsxToAnd',p);
                         RemoveInstruction(hp1);
                         Exit;
                       end;
                  end
            else if (taicpu(p).oper[1]^.typ = top_reg) and
              (hp1.typ = ait_instruction) and
              (taicpu(hp1).is_jmp) and
              (taicpu(hp1).opcode<>A_JMP) and
              not(RegInUsedRegs(taicpu(p).oper[1]^.reg,UsedRegs)) then
              begin
                { change
                    and x, reg
                    jxx
                  to
                    test x, reg
                    jxx
                  if reg is deallocated before the
                  jump, but only if it's a conditional jump (PFV)
                }
                taicpu(p).opcode := A_TEST;
                Exit;
              end;
          end;

        { Lone AND tests }
        if MatchOpType(taicpu(p),top_const,top_reg) then
          begin
            {
              - Convert and $0xFF,reg to and reg,reg if reg is 8-bit
              - Convert and $0xFFFF,reg to and reg,reg if reg is 16-bit
              - Convert and $0xFFFFFFFF,reg to and reg,reg if reg is 32-bit
            }
            if ((taicpu(p).oper[0]^.val = $FF) and (taicpu(p).opsize = S_B)) or
              ((taicpu(p).oper[0]^.val = $FFFF) and (taicpu(p).opsize = S_W)) or
              ((taicpu(p).oper[0]^.val = $FFFFFFFF) and (taicpu(p).opsize = S_L)) then
              begin
                taicpu(p).loadreg(0, taicpu(p).oper[1]^.reg);
                if taicpu(p).opsize = S_L then
                  begin
                    Include(OptsToCheck,aoc_MovAnd2Mov_3);
                    Result := True;
                  end;
              end;
          end;

      end;


    function TX86AsmOptimizer.OptPass2Lea(var p : tai) : Boolean;
      begin
        Result:=false;
        if not (RegInUsedRegs(NR_DEFAULTFLAGS,UsedRegs)) and
          MatchReference(taicpu(p).oper[0]^.ref^,taicpu(p).oper[1]^.reg,NR_INVALID) and
          (taicpu(p).oper[0]^.ref^.index<>NR_NO) then
          begin
            taicpu(p).loadreg(1,taicpu(p).oper[0]^.ref^.base);
            taicpu(p).loadreg(0,taicpu(p).oper[0]^.ref^.index);
            taicpu(p).opcode:=A_ADD;
            DebugMsg(SPeepholeOptimization + 'Lea2AddBase done',p);
            result:=true;
          end

        else if not (RegInUsedRegs(NR_DEFAULTFLAGS,UsedRegs)) and
          MatchReference(taicpu(p).oper[0]^.ref^,NR_INVALID,taicpu(p).oper[1]^.reg) and
          (taicpu(p).oper[0]^.ref^.base<>NR_NO) then
          begin
            taicpu(p).loadreg(1,taicpu(p).oper[0]^.ref^.index);
            taicpu(p).loadreg(0,taicpu(p).oper[0]^.ref^.base);
            taicpu(p).opcode:=A_ADD;
            DebugMsg(SPeepholeOptimization + 'Lea2AddIndex done',p);
            result:=true;
          end;
      end;


    function TX86AsmOptimizer.OptPass2SUB(var p: tai): Boolean;
      var
        hp1: tai; NewRef: TReference;
      begin
        { Change:
            subl/q $x,%reg1
            movl/q %reg1,%reg2
          To:
            leal/q $-x(%reg1),%reg2
            subl/q $x,%reg1

          Breaks the dependency chain and potentially permits the removal of
          a CMP instruction if one follows.
        }
        Result := False;
        if not (cs_opt_size in current_settings.optimizerswitches) and
          (taicpu(p).opsize in [S_L{$ifdef x86_64}, S_Q{$endif x86_64}]) and
          MatchOpType(taicpu(p),top_const,top_reg) and
          GetNextInstruction(p, hp1) and
          MatchInstruction(hp1, A_MOV, [taicpu(p).opsize]) and
          (taicpu(hp1).oper[1]^.typ = top_reg) and
          MatchOperand(taicpu(hp1).oper[0]^, taicpu(p).oper[1]^.reg) then
          begin
            { Change the MOV instruction to a LEA instruction, and update the
              first operand }
            reference_reset(NewRef, 1, []);
            NewRef.base := taicpu(p).oper[1]^.reg;
            NewRef.scalefactor := 1;
            NewRef.offset := -taicpu(p).oper[0]^.val;

            taicpu(hp1).opcode := A_LEA;
            taicpu(hp1).loadref(0, NewRef);

            { Move what is now the LEA instruction to before the SUB instruction }
            Asml.Remove(hp1);
            Asml.InsertBefore(hp1, p);
            AllocRegBetween(taicpu(hp1).oper[1]^.reg, hp1, p, UsedRegs);

            DebugMsg(SPeepholeOptimization + 'SubMov2LeaSub', p);
            Result := True;
          end;
      end;


    function TX86AsmOptimizer.SkipSimpleInstructions(var hp1 : tai) : Boolean;
      begin
        { we can skip all instructions not messing with the stack pointer }
        while assigned(hp1) and {MatchInstruction(taicpu(hp1),[A_LEA,A_MOV,A_MOVQ,A_MOVSQ,A_MOVSX,A_MOVSXD,A_MOVZX,
          A_AND,A_OR,A_XOR,A_ADD,A_SHR,A_SHL,A_IMUL,A_SETcc,A_SAR,A_SUB,A_TEST,A_CMOVcc,
          A_MOVSS,A_MOVSD,A_MOVAPS,A_MOVUPD,A_MOVAPD,A_MOVUPS,
          A_VMOVSS,A_VMOVSD,A_VMOVAPS,A_VMOVUPD,A_VMOVAPD,A_VMOVUPS],[]) and}
          ({(taicpu(hp1).ops=0) or }
           ({(MatchOpType(taicpu(hp1),top_reg,top_reg) or MatchOpType(taicpu(hp1),top_const,top_reg) or
             (MatchOpType(taicpu(hp1),top_ref,top_reg))
            ) and }
            not(RegInInstruction(NR_STACK_POINTER_REG,hp1)) { and not(RegInInstruction(NR_FRAME_POINTER_REG,hp1))}
           )
          ) do
          GetNextInstruction(hp1,hp1);
        Result:=assigned(hp1);
      end;


    function TX86AsmOptimizer.PostPeepholeOptLea(var p : tai) : Boolean;

      var
        hp1, hp2, hp3, hp4: tai;
      begin
        Result:=false;
        { replace
            leal(q) x(<stackpointer>),<stackpointer>
            call   procname
            leal(q) -x(<stackpointer>),<stackpointer>
            ret
          by
            jmp    procname

          but do it only on level 4 because it destroys stack back traces
        }
        if (cs_opt_level4 in current_settings.optimizerswitches) and
          MatchOpType(taicpu(p),top_ref,top_reg) and
          (taicpu(p).oper[0]^.ref^.base=NR_STACK_POINTER_REG) and
          (taicpu(p).oper[0]^.ref^.index=NR_NO) and
          { the -8 or -24 are not required, but bail out early if possible,
            higher values are unlikely }
          ((taicpu(p).oper[0]^.ref^.offset=-8) or
           (taicpu(p).oper[0]^.ref^.offset=-24))  and
          (taicpu(p).oper[0]^.ref^.symbol=nil) and
          (taicpu(p).oper[0]^.ref^.relsymbol=nil) and
          (taicpu(p).oper[0]^.ref^.segment=NR_NO) and
          (taicpu(p).oper[1]^.reg=NR_STACK_POINTER_REG) and
          GetNextInstruction(p, hp1) and
          { Take a copy of hp1 }
          SetAndTest(hp1, hp4) and
          { trick to skip label }
          ((hp1.typ=ait_instruction) or GetNextInstruction(hp1, hp1)) and
          SkipSimpleInstructions(hp1) and
          MatchInstruction(hp1,A_CALL,[S_NO]) and
          GetNextInstruction(hp1, hp2) and
          MatchInstruction(hp2,A_LEA,[taicpu(p).opsize]) and
          MatchOpType(taicpu(hp2),top_ref,top_reg) and
          (taicpu(hp2).oper[0]^.ref^.offset=-taicpu(p).oper[0]^.ref^.offset) and
          (taicpu(hp2).oper[0]^.ref^.base=NR_STACK_POINTER_REG) and
          (taicpu(hp2).oper[0]^.ref^.index=NR_NO) and
          (taicpu(hp2).oper[0]^.ref^.symbol=nil) and
          (taicpu(hp2).oper[0]^.ref^.relsymbol=nil) and
          (taicpu(hp2).oper[0]^.ref^.segment=NR_NO) and
          (taicpu(hp2).oper[1]^.reg=NR_STACK_POINTER_REG) and
          GetNextInstruction(hp2, hp3) and
          { trick to skip label }
          ((hp3.typ=ait_instruction) or GetNextInstruction(hp3, hp3)) and
          MatchInstruction(hp3,A_RET,[S_NO]) and
          (taicpu(hp3).ops=0) then
          begin
            taicpu(hp1).opcode := A_JMP;
            taicpu(hp1).is_jmp := true;
            DebugMsg(SPeepholeOptimization + 'LeaCallLeaRet2Jmp done',p);
            RemoveCurrentP(p, hp4);
            RemoveInstruction(hp2);
            RemoveInstruction(hp3);
            Result:=true;
          end;
      end;


    function TX86AsmOptimizer.PostPeepholeOptPush(var p : tai) : Boolean;

      var
        hp1, hp2, hp3, hp4: tai;
      begin
        Result:=false;
{$ifdef x86_64}
        { replace
            push %rax
            call   procname
            pop %rcx
            ret
          by
            jmp    procname

          but do it only on level 4 because it destroys stack back traces

          It depends on the fact, that the sequence push rax/pop rcx is used for stack alignment as rcx is volatile
          for all supported calling conventions
        }
        if (cs_opt_level4 in current_settings.optimizerswitches) and
          MatchOpType(taicpu(p),top_reg) and
          (taicpu(p).oper[0]^.reg=NR_RAX) and
          GetNextInstruction(p, hp1) and
          { Take a copy of hp1 }
          SetAndTest(hp1, hp4) and
          { trick to skip label }
          ((hp1.typ=ait_instruction) or GetNextInstruction(hp1, hp1)) and
          SkipSimpleInstructions(hp1) and
          MatchInstruction(hp1,A_CALL,[S_NO]) and
          GetNextInstruction(hp1, hp2) and
          MatchInstruction(hp2,A_POP,[taicpu(p).opsize]) and
          MatchOpType(taicpu(hp2),top_reg) and
          (taicpu(hp2).oper[0]^.reg=NR_RCX) and
          GetNextInstruction(hp2, hp3) and
          { trick to skip label }
          ((hp3.typ=ait_instruction) or GetNextInstruction(hp3, hp3)) and
          MatchInstruction(hp3,A_RET,[S_NO]) and
          (taicpu(hp3).ops=0) then
          begin
            taicpu(hp1).opcode := A_JMP;
            taicpu(hp1).is_jmp := true;
            DebugMsg(SPeepholeOptimization + 'PushCallPushRet2Jmp done',p);
            RemoveCurrentP(p, hp4);
            RemoveInstruction(hp2);
            RemoveInstruction(hp3);
            Result:=true;
          end;
{$endif x86_64}
      end;


    function TX86AsmOptimizer.PostPeepholeOptMov(var p : tai) : Boolean;
      var
        Value, RegName: string;
      begin
        Result:=false;
        if (taicpu(p).oper[1]^.typ = top_reg) and (taicpu(p).oper[0]^.typ = top_const) then
          begin

            case taicpu(p).oper[0]^.val of
            0:
              { Don't make this optimisation if the CPU flags are required, since XOR scrambles them }
              if not (RegInUsedRegs(NR_DEFAULTFLAGS,UsedRegs)) then
                begin
                  { change "mov $0,%reg" into "xor %reg,%reg" }
                  taicpu(p).opcode := A_XOR;
                  taicpu(p).loadReg(0,taicpu(p).oper[1]^.reg);
                  Result := True;
                end;
            $1..$FFFFFFFF:
              begin
                { Code size reduction by J. Gareth "Kit" Moreton }
                { change 64-bit register to 32-bit register to reduce code size (upper 32 bits will be set to zero) }
                case taicpu(p).opsize of
                  S_Q:
                    begin
                      RegName := debug_regname(taicpu(p).oper[1]^.reg); { 64-bit register name }
                      Value := debug_tostr(taicpu(p).oper[0]^.val);

                      { The actual optimization }
                      setsubreg(taicpu(p).oper[1]^.reg, R_SUBD);
                      taicpu(p).changeopsize(S_L);

                      DebugMsg(SPeepholeOptimization + 'movq $' + Value + ',' + RegName + ' -> movl $' + Value + ',' + debug_regname(taicpu(p).oper[1]^.reg) + ' (immediate can be represented with just 32 bits)', p);
                      Result := True;
                    end;
                  else
                    { Do nothing };
                end;
              end;
            -1:
              { Don't make this optimisation if the CPU flags are required, since OR scrambles them }
              if (cs_opt_size in current_settings.optimizerswitches) and
                (taicpu(p).opsize <> S_B) and
                not (RegInUsedRegs(NR_DEFAULTFLAGS,UsedRegs)) then
                begin
                  { change "mov $-1,%reg" into "or $-1,%reg" }
                  { NOTES:
                    - No size saving is made when changing a Word-sized assignment unless the register is AX (smaller encoding)
                    - This operation creates a false dependency on the register, so only do it when optimising for size
                    - It is possible to set memory operands using this method, but this creates an even greater false dependency, so don't do this at all
                  }
                  taicpu(p).opcode := A_OR;
                  Result := True;
                end;
            end;
          end;
      end;


    function TX86AsmOptimizer.PostPeepholeOptMOVSX(var p : tai) : boolean;
      begin
        Result := False;
        if not MatchOpType(taicpu(p), top_reg, top_reg) then
          Exit;

        { Convert:
            movswl %ax,%eax  -> cwtl
            movslq %eax,%rax -> cdqe

            NOTE: Don't convert movswl %al,%ax to cbw, because cbw and cwde
              refer to the same opcode and depends only on the assembler's
              current operand-size attribute. [Kit]
        }
        with taicpu(p) do
          case opsize of
            S_WL:
              if (oper[0]^.reg = NR_AX) and (oper[1]^.reg = NR_EAX) then
                begin
                  DebugMsg(SPeepholeOptimization + 'Converted movswl %ax,%eax to cwtl', p);
                  opcode := A_CWDE;
                  clearop(0);
                  clearop(1);
                  ops := 0;
                  Result := True;
                end;
{$ifdef x86_64}
            S_LQ:
              if (oper[0]^.reg = NR_EAX) and (oper[1]^.reg = NR_RAX) then
                begin
                  DebugMsg(SPeepholeOptimization + 'Converted movslq %eax,%rax to cltq', p);
                  opcode := A_CDQE;
                  clearop(0);
                  clearop(1);
                  ops := 0;
                  Result := True;
                end;
{$endif x86_64}
            else
              ;
          end;
      end;


    function TX86AsmOptimizer.PostPeepholeOptCmp(var p : tai) : Boolean;
      begin
        Result:=false;
        { change "cmp $0, %reg" to "test %reg, %reg" }
        if MatchOpType(taicpu(p),top_const,top_reg) and
           (taicpu(p).oper[0]^.val = 0) then
          begin
            taicpu(p).opcode := A_TEST;
            taicpu(p).loadreg(0,taicpu(p).oper[1]^.reg);
            Result:=true;
          end;
      end;


    function TX86AsmOptimizer.PostPeepholeOptTestOr(var p : tai) : Boolean;
      var
        IsTestConstX : Boolean;
        hp1,hp2 : tai;
      begin
        Result:=false;
        { removes the line marked with (x) from the sequence
          and/or/xor/add/sub/... $x, %y
          test/or %y, %y  | test $-1, %y    (x)
          j(n)z _Label
             as the first instruction already adjusts the ZF
             %y operand may also be a reference }
        IsTestConstX:=(taicpu(p).opcode=A_TEST) and
          MatchOperand(taicpu(p).oper[0]^,-1);
        if (OpsEqual(taicpu(p).oper[0]^,taicpu(p).oper[1]^) or IsTestConstX) and
           GetLastInstruction(p, hp1) and
           (tai(hp1).typ = ait_instruction) and
           GetNextInstruction(p,hp2) and
           MatchInstruction(hp2,A_SETcc,A_Jcc,A_CMOVcc,[]) then
          case taicpu(hp1).opcode Of
            A_ADD, A_SUB, A_OR, A_XOR, A_AND:
              begin
                if OpsEqual(taicpu(hp1).oper[1]^,taicpu(p).oper[1]^) and
                  { does not work in case of overflow for G(E)/L(E)/C_O/C_NO }
                  { and in case of carry for A(E)/B(E)/C/NC                  }
                   ((taicpu(hp2).condition in [C_Z,C_NZ,C_E,C_NE]) or
                    ((taicpu(hp1).opcode <> A_ADD) and
                     (taicpu(hp1).opcode <> A_SUB))) then
                  begin
                    RemoveCurrentP(p, hp2);
                    Result:=true;
                  end;
              end;
            A_SHL, A_SAL, A_SHR, A_SAR:
              begin
                if OpsEqual(taicpu(hp1).oper[1]^,taicpu(p).oper[1]^) and
                  { SHL/SAL/SHR/SAR with a value of 0 do not change the flags }
                  { therefore, it's only safe to do this optimization for     }
                  { shifts by a (nonzero) constant                            }
                   (taicpu(hp1).oper[0]^.typ = top_const) and
                   (taicpu(hp1).oper[0]^.val <> 0) and
                  { does not work in case of overflow for G(E)/L(E)/C_O/C_NO }
                  { and in case of carry for A(E)/B(E)/C/NC                  }
                   (taicpu(hp2).condition in [C_Z,C_NZ,C_E,C_NE]) then
                  begin
                    RemoveCurrentP(p, hp2);
                    Result:=true;
                  end;
              end;
            A_DEC, A_INC, A_NEG:
              begin
                if OpsEqual(taicpu(hp1).oper[0]^,taicpu(p).oper[1]^) and
                  { does not work in case of overflow for G(E)/L(E)/C_O/C_NO }
                  { and in case of carry for A(E)/B(E)/C/NC                  }
                  (taicpu(hp2).condition in [C_Z,C_NZ,C_E,C_NE]) then
                  begin
                    case taicpu(hp1).opcode of
                      A_DEC, A_INC:
                        { replace inc/dec with add/sub 1, because inc/dec doesn't set the carry flag }
                        begin
                          case taicpu(hp1).opcode Of
                            A_DEC: taicpu(hp1).opcode := A_SUB;
                            A_INC: taicpu(hp1).opcode := A_ADD;
                            else
                              ;
                          end;
                          taicpu(hp1).loadoper(1,taicpu(hp1).oper[0]^);
                          taicpu(hp1).loadConst(0,1);
                          taicpu(hp1).ops:=2;
                        end;
                      else
                        ;
                    end;
                    RemoveCurrentP(p, hp2);
                    Result:=true;
                  end;
              end
          else
            { change "test  $-1,%reg" into "test %reg,%reg" }
            if IsTestConstX and (taicpu(p).oper[1]^.typ=top_reg) then
              taicpu(p).loadoper(0,taicpu(p).oper[1]^);
          end { case }
        { change "test  $-1,%reg" into "test %reg,%reg" }
        else if IsTestConstX and (taicpu(p).oper[1]^.typ=top_reg) then
          taicpu(p).loadoper(0,taicpu(p).oper[1]^);
      end;


    function TX86AsmOptimizer.PostPeepholeOptCall(var p : tai) : Boolean;
      var
        hp1 : tai;
{$ifndef x86_64}
        hp2 : taicpu;
{$endif x86_64}
      begin
        Result:=false;
{$ifndef x86_64}
        { don't do this on modern CPUs, this really hurts them due to
          broken call/ret pairing }
        if (current_settings.optimizecputype < cpu_Pentium2) and
           not(cs_create_pic in current_settings.moduleswitches) and
           GetNextInstruction(p, hp1) and
           MatchInstruction(hp1,A_JMP,[S_NO]) and
           MatchOpType(taicpu(hp1),top_ref) and
           (taicpu(hp1).oper[0]^.ref^.refaddr=addr_full) then
          begin
            hp2 := taicpu.Op_sym(A_PUSH,S_L,taicpu(hp1).oper[0]^.ref^.symbol);
            InsertLLItem(p.previous, p, hp2);
            taicpu(p).opcode := A_JMP;
            taicpu(p).is_jmp := true;
            RemoveInstruction(hp1);
            Result:=true;
          end
        else
{$endif x86_64}
        { replace
            call   procname
            ret
          by
            jmp    procname

          but do it only on level 4 because it destroys stack back traces

          else if the subroutine is marked as no return, remove the ret
        }
        if ((cs_opt_level4 in current_settings.optimizerswitches) or
          (po_noreturn in current_procinfo.procdef.procoptions)) and
          GetNextInstruction(p, hp1) and
          MatchInstruction(hp1,A_RET,[S_NO]) and
          (taicpu(hp1).ops=0) then
          begin
            if (cs_opt_level4 in current_settings.optimizerswitches) and
              { we might destroy stack alignment here if we do not do a call }
              (target_info.stackalign<=sizeof(SizeUInt)) then
              begin
                taicpu(p).opcode := A_JMP;
                taicpu(p).is_jmp := true;
                DebugMsg(SPeepholeOptimization + 'CallRet2Jmp done',p);
              end
            else
              DebugMsg(SPeepholeOptimization + 'CallRet2Call done',p);
            RemoveInstruction(hp1);
            Result:=true;
          end;
      end;


{$ifdef x86_64}
    function TX86AsmOptimizer.PostPeepholeOptMovzx(var p : tai) : Boolean;
      var
        PreMessage: string;
      begin
        Result := False;
        { Code size reduction by J. Gareth "Kit" Moreton }
        { Convert MOVZBQ and MOVZWQ to MOVZBL and MOVZWL respectively if it removes the REX prefix }
        if (taicpu(p).opsize in [S_BQ, S_WQ]) and
          (getsupreg(taicpu(p).oper[1]^.reg) in [RS_RAX, RS_RCX, RS_RDX, RS_RBX, RS_RSI, RS_RDI, RS_RBP, RS_RSP])
        then
          begin
            { Has 64-bit register name and opcode suffix }
            PreMessage := 'movz' + debug_opsize2str(taicpu(p).opsize) + ' ' + debug_operstr(taicpu(p).oper[0]^) + ',' + debug_regname(taicpu(p).oper[1]^.reg) + ' -> movz';

            { The actual optimization }
            setsubreg(taicpu(p).oper[1]^.reg, R_SUBD);
            if taicpu(p).opsize = S_BQ then
              taicpu(p).changeopsize(S_BL)
            else
              taicpu(p).changeopsize(S_WL);

            DebugMsg(SPeepholeOptimization + PreMessage +
              debug_opsize2str(taicpu(p).opsize) + ' ' + debug_operstr(taicpu(p).oper[0]^) + ',' + debug_regname(taicpu(p).oper[1]^.reg) + ' (removes REX prefix)', p);
          end;
      end;


    function TX86AsmOptimizer.PostPeepholeOptXor(var p : tai) : Boolean;
      var
        PreMessage, RegName: string;
      begin
        { Code size reduction by J. Gareth "Kit" Moreton }
        { change "xorq %reg,%reg" to "xorl %reg,%reg" for %rax, %rcx, %rdx, %rbx, %rsi, %rdi, %rbp and %rsp,
          as this removes the REX prefix }

        Result := False;
        if not OpsEqual(taicpu(p).oper[0]^,taicpu(p).oper[1]^) then
          Exit;

        if taicpu(p).oper[0]^.typ <> top_reg then
          { Should be impossible if both operands were equal, since one of XOR's operands must be a register }
          InternalError(2018011500);

        case taicpu(p).opsize of
          S_Q:
            begin
              if (getsupreg(taicpu(p).oper[0]^.reg) in [RS_RAX, RS_RCX, RS_RDX, RS_RBX, RS_RSI, RS_RDI, RS_RBP, RS_RSP]) then
                begin
                  RegName := debug_regname(taicpu(p).oper[0]^.reg); { 64-bit register name }
                  PreMessage := 'xorq ' + RegName + ',' + RegName + ' -> xorl ';

                  { The actual optimization }
                  setsubreg(taicpu(p).oper[0]^.reg, R_SUBD);
                  setsubreg(taicpu(p).oper[1]^.reg, R_SUBD);
                  taicpu(p).changeopsize(S_L);

                  RegName := debug_regname(taicpu(p).oper[0]^.reg); { 32-bit register name }

                  DebugMsg(SPeepholeOptimization + PreMessage + RegName + ',' + RegName + ' (removes REX prefix)', p);
                end;
            end;
          else
            ;
        end;
      end;
{$endif}


    class procedure TX86AsmOptimizer.OptimizeRefs(var p: taicpu);
      var
        OperIdx: Integer;
      begin
        for OperIdx := 0 to p.ops - 1 do
          if p.oper[OperIdx]^.typ = top_ref then
            optimize_ref(p.oper[OperIdx]^.ref^, False);
      end;

end.