ARMInstrInfo.td

//===----------------------------------------------------------------------===//
//  Multiply Instructions.
//

let isCommutable = 1 in
def MUL   : AsMul1I<0b0000000, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
                    "mul", " $dst, $a, $b",
                   [(set GPR:$dst, (mul GPR:$a, GPR:$b))]>;

def MLA   : AsMul1I<0b0000001, (outs GPR:$dst), (ins GPR:$a, GPR:$b, GPR:$c),
                    "mla", " $dst, $a, $b, $c",
                   [(set GPR:$dst, (add (mul GPR:$a, GPR:$b), GPR:$c))]>;

// Extra precision multiplies with low / high results
let neverHasSideEffects = 1 in {
let isCommutable = 1 in {
def SMULL : AsMul1I<0b0000110, (outs GPR:$ldst, GPR:$hdst),
                               (ins GPR:$a, GPR:$b),
                    "smull", " $ldst, $hdst, $a, $b", []>;

def UMULL : AsMul1I<0b0000100, (outs GPR:$ldst, GPR:$hdst),
                               (ins GPR:$a, GPR:$b),
                    "umull", " $ldst, $hdst, $a, $b", []>;
}

// Multiply + accumulate
def SMLAL : AsMul1I<0b0000111, (outs GPR:$ldst, GPR:$hdst),
                               (ins GPR:$a, GPR:$b),
                    "smlal", " $ldst, $hdst, $a, $b", []>;

def UMLAL : AsMul1I<0b0000101, (outs GPR:$ldst, GPR:$hdst),
                               (ins GPR:$a, GPR:$b),
                    "umlal", " $ldst, $hdst, $a, $b", []>;

def UMAAL : AMul1I <0b0000010, (outs GPR:$ldst, GPR:$hdst),
                               (ins GPR:$a, GPR:$b),
                    "umaal", " $ldst, $hdst, $a, $b", []>,
                    Requires<[IsARM, HasV6]>;
} // neverHasSideEffects

// Most significant word multiply
def SMMUL : AMul2I <0b0111010, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
               "smmul", " $dst, $a, $b",
               [(set GPR:$dst, (mulhs GPR:$a, GPR:$b))]>,
            Requires<[IsARM, HasV6]> {
  let Inst{7-4}   = 0b0001;
  let Inst{15-12} = 0b1111;
}

def SMMLA : AMul2I <0b0111010, (outs GPR:$dst), (ins GPR:$a, GPR:$b, GPR:$c),
               "smmla", " $dst, $a, $b, $c",
               [(set GPR:$dst, (add (mulhs GPR:$a, GPR:$b), GPR:$c))]>,
            Requires<[IsARM, HasV6]> {
  let Inst{7-4}   = 0b0001;
}


def SMMLS : AMul2I <0b0111010, (outs GPR:$dst), (ins GPR:$a, GPR:$b, GPR:$c),
               "smmls", " $dst, $a, $b, $c",
               [(set GPR:$dst, (sub GPR:$c, (mulhs GPR:$a, GPR:$b)))]>,
            Requires<[IsARM, HasV6]> {
  let Inst{7-4}   = 0b1101;
}

multiclass AI_smul<string opc, PatFrag opnode> {
  def BB : AMulxyI<0b0001011, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
              !strconcat(opc, "bb"), " $dst, $a, $b",
              [(set GPR:$dst, (opnode (sext_inreg GPR:$a, i16),
                                      (sext_inreg GPR:$b, i16)))]>,
           Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 0;
             let Inst{6} = 0;
           }

  def BT : AMulxyI<0b0001011, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
              !strconcat(opc, "bt"), " $dst, $a, $b",
              [(set GPR:$dst, (opnode (sext_inreg GPR:$a, i16),
                                      (sra GPR:$b, (i32 16))))]>,
           Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 0;
             let Inst{6} = 1;
           }

  def TB : AMulxyI<0b0001011, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
              !strconcat(opc, "tb"), " $dst, $a, $b",
              [(set GPR:$dst, (opnode (sra GPR:$a, (i32 16)),
                                      (sext_inreg GPR:$b, i16)))]>,
           Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 1;
             let Inst{6} = 0;
           }

  def TT : AMulxyI<0b0001011, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
              !strconcat(opc, "tt"), " $dst, $a, $b",
              [(set GPR:$dst, (opnode (sra GPR:$a, (i32 16)),
                                      (sra GPR:$b, (i32 16))))]>,
            Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 1;
             let Inst{6} = 1;
           }

  def WB : AMulxyI<0b0001001, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
              !strconcat(opc, "wb"), " $dst, $a, $b",
              [(set GPR:$dst, (sra (opnode GPR:$a,
                                    (sext_inreg GPR:$b, i16)), (i32 16)))]>,
           Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 1;
             let Inst{6} = 0;
           }

  def WT : AMulxyI<0b0001001, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
              !strconcat(opc, "wt"), " $dst, $a, $b",
              [(set GPR:$dst, (sra (opnode GPR:$a,
                                    (sra GPR:$b, (i32 16))), (i32 16)))]>,
            Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 1;
             let Inst{6} = 1;
           }
}


multiclass AI_smla<string opc, PatFrag opnode> {
  def BB : AMulxyI<0b0001000, (outs GPR:$dst), (ins GPR:$a, GPR:$b, GPR:$acc),
              !strconcat(opc, "bb"), " $dst, $a, $b, $acc",
              [(set GPR:$dst, (add GPR:$acc,
                               (opnode (sext_inreg GPR:$a, i16),
                                       (sext_inreg GPR:$b, i16))))]>,
           Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 0;
             let Inst{6} = 0;
           }

  def BT : AMulxyI<0b0001000, (outs GPR:$dst), (ins GPR:$a, GPR:$b, GPR:$acc),
              !strconcat(opc, "bt"), " $dst, $a, $b, $acc",
              [(set GPR:$dst, (add GPR:$acc, (opnode (sext_inreg GPR:$a, i16),
                                                     (sra GPR:$b, (i32 16)))))]>,
           Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 0;
             let Inst{6} = 1;
           }

  def TB : AMulxyI<0b0001000, (outs GPR:$dst), (ins GPR:$a, GPR:$b, GPR:$acc),
              !strconcat(opc, "tb"), " $dst, $a, $b, $acc",
              [(set GPR:$dst, (add GPR:$acc, (opnode (sra GPR:$a, (i32 16)),
                                                 (sext_inreg GPR:$b, i16))))]>,
           Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 1;
             let Inst{6} = 0;
           }

  def TT : AMulxyI<0b0001000, (outs GPR:$dst), (ins GPR:$a, GPR:$b, GPR:$acc),
              !strconcat(opc, "tt"), " $dst, $a, $b, $acc",
              [(set GPR:$dst, (add GPR:$acc, (opnode (sra GPR:$a, (i32 16)),
                                                     (sra GPR:$b, (i32 16)))))]>,
            Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 1;
             let Inst{6} = 1;
           }

  def WB : AMulxyI<0b0001001, (outs GPR:$dst), (ins GPR:$a, GPR:$b, GPR:$acc),
              !strconcat(opc, "wb"), " $dst, $a, $b, $acc",
              [(set GPR:$dst, (add GPR:$acc, (sra (opnode GPR:$a,
                                       (sext_inreg GPR:$b, i16)), (i32 16))))]>,
           Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 0;
             let Inst{6} = 0;
           }

  def WT : AMulxyI<0b0001001, (outs GPR:$dst), (ins GPR:$a, GPR:$b, GPR:$acc),
              !strconcat(opc, "wt"), " $dst, $a, $b, $acc",
              [(set GPR:$dst, (add GPR:$acc, (sra (opnode GPR:$a,
                                         (sra GPR:$b, (i32 16))), (i32 16))))]>,
            Requires<[IsARM, HasV5TE]> {
             let Inst{5} = 0;
             let Inst{6} = 1;
           }
}

defm SMUL : AI_smul<"smul", BinOpFrag<(mul node:$LHS, node:$RHS)>>;
defm SMLA : AI_smla<"smla", BinOpFrag<(mul node:$LHS, node:$RHS)>>;

// TODO: Halfword multiple accumulate long: SMLAL<x><y>
// TODO: Dual halfword multiple: SMUAD, SMUSD, SMLAD, SMLSD, SMLALD, SMLSLD

//===----------------------------------------------------------------------===//
//  Misc. Arithmetic Instructions.
//

def CLZ  : AMiscA1I<0b000010110, (outs GPR:$dst), (ins GPR:$src),
              "clz", " $dst, $src",
              [(set GPR:$dst, (ctlz GPR:$src))]>, Requires<[IsARM, HasV5T]> {
  let Inst{7-4}   = 0b0001;
  let Inst{11-8}  = 0b1111;
  let Inst{19-16} = 0b1111;
}

def REV  : AMiscA1I<0b01101011, (outs GPR:$dst), (ins GPR:$src),
              "rev", " $dst, $src",
              [(set GPR:$dst, (bswap GPR:$src))]>, Requires<[IsARM, HasV6]> {
  let Inst{7-4}   = 0b0011;
  let Inst{11-8}  = 0b1111;
  let Inst{19-16} = 0b1111;
}

def REV16 : AMiscA1I<0b01101011, (outs GPR:$dst), (ins GPR:$src),
               "rev16", " $dst, $src",
               [(set GPR:$dst,
                   (or (and (srl GPR:$src, (i32 8)), 0xFF),
                       (or (and (shl GPR:$src, (i32 8)), 0xFF00),
                           (or (and (srl GPR:$src, (i32 8)), 0xFF0000),
                               (and (shl GPR:$src, (i32 8)), 0xFF000000)))))]>,
               Requires<[IsARM, HasV6]> {
  let Inst{7-4}   = 0b1011;
  let Inst{11-8}  = 0b1111;
  let Inst{19-16} = 0b1111;
}

def REVSH : AMiscA1I<0b01101111, (outs GPR:$dst), (ins GPR:$src),
               "revsh", " $dst, $src",
               [(set GPR:$dst,
                  (sext_inreg
                    (or (srl (and GPR:$src, 0xFF00), (i32 8)),
                        (shl GPR:$src, (i32 8))), i16))]>,
               Requires<[IsARM, HasV6]> {
  let Inst{7-4}   = 0b1011;
  let Inst{11-8}  = 0b1111;
  let Inst{19-16} = 0b1111;
}

def PKHBT : AMiscA1I<0b01101000, (outs GPR:$dst),
                                 (ins GPR:$src1, GPR:$src2, i32imm:$shamt),
               "pkhbt", " $dst, $src1, $src2, LSL $shamt",
               [(set GPR:$dst, (or (and GPR:$src1, 0xFFFF),
                                   (and (shl GPR:$src2, (i32 imm:$shamt)),
                                        0xFFFF0000)))]>,
               Requires<[IsARM, HasV6]> {
  let Inst{6-4} = 0b001;
}

// Alternate cases for PKHBT where identities eliminate some nodes.
def : ARMV6Pat<(or (and GPR:$src1, 0xFFFF), (and GPR:$src2, 0xFFFF0000)),
               (PKHBT GPR:$src1, GPR:$src2, 0)>;
def : ARMV6Pat<(or (and GPR:$src1, 0xFFFF), (shl GPR:$src2, imm16_31:$shamt)),
               (PKHBT GPR:$src1, GPR:$src2, imm16_31:$shamt)>;


def PKHTB : AMiscA1I<0b01101000, (outs GPR:$dst),
                                 (ins GPR:$src1, GPR:$src2, i32imm:$shamt),
               "pkhtb", " $dst, $src1, $src2, ASR $shamt",
               [(set GPR:$dst, (or (and GPR:$src1, 0xFFFF0000),
                                   (and (sra GPR:$src2, imm16_31:$shamt),
                                        0xFFFF)))]>, Requires<[IsARM, HasV6]> {
  let Inst{6-4} = 0b101;
}

// Alternate cases for PKHTB where identities eliminate some nodes.  Note that
// a shift amount of 0 is *not legal* here, it is PKHBT instead.
def : ARMV6Pat<(or (and GPR:$src1, 0xFFFF0000), (srl GPR:$src2, (i32 16))),
               (PKHTB GPR:$src1, GPR:$src2, 16)>;
def : ARMV6Pat<(or (and GPR:$src1, 0xFFFF0000),
                   (and (srl GPR:$src2, imm1_15:$shamt), 0xFFFF)),
               (PKHTB GPR:$src1, GPR:$src2, imm1_15:$shamt)>;

//===----------------------------------------------------------------------===//
//  Comparison Instructions...
//

defm CMP  : AI1_cmp_irs<0b1010, "cmp",
                        BinOpFrag<(ARMcmp node:$LHS, node:$RHS)>>;
defm CMN  : AI1_cmp_irs<0b1011, "cmn",
                        BinOpFrag<(ARMcmp node:$LHS,(ineg node:$RHS))>>;

// Note that TST/TEQ don't set all the same flags that CMP does!
defm TST  : AI1_cmp_irs<0b1000, "tst",
                        BinOpFrag<(ARMcmpZ (and node:$LHS, node:$RHS), 0)>, 1>;
defm TEQ  : AI1_cmp_irs<0b1001, "teq",
                        BinOpFrag<(ARMcmpZ (xor node:$LHS, node:$RHS), 0)>, 1>;

defm CMPz  : AI1_cmp_irs<0b1010, "cmp",
                         BinOpFrag<(ARMcmpZ node:$LHS, node:$RHS)>>;
defm CMNz  : AI1_cmp_irs<0b1011, "cmn",
                         BinOpFrag<(ARMcmpZ node:$LHS,(ineg node:$RHS))>>;

def : ARMPat<(ARMcmp GPR:$src, so_imm_neg:$imm),
             (CMNri  GPR:$src, so_imm_neg:$imm)>;

def : ARMPat<(ARMcmpZ GPR:$src, so_imm_neg:$imm),
             (CMNri  GPR:$src, so_imm_neg:$imm)>;


// Conditional moves
// FIXME: should be able to write a pattern for ARMcmov, but can't use
// a two-value operand where a dag node expects two operands. :( 
def MOVCCr : AI1<0b1101, (outs GPR:$dst), (ins GPR:$false, GPR:$true), DPFrm,
                "mov", " $dst, $true",
      [/*(set GPR:$dst, (ARMcmov GPR:$false, GPR:$true, imm:$cc, CCR:$ccr))*/]>,
                RegConstraint<"$false = $dst">, UnaryDP;

def MOVCCs : AI1<0b1101, (outs GPR:$dst),
                        (ins GPR:$false, so_reg:$true), DPSoRegFrm,
                "mov", " $dst, $true",
   [/*(set GPR:$dst, (ARMcmov GPR:$false, so_reg:$true, imm:$cc, CCR:$ccr))*/]>,
                RegConstraint<"$false = $dst">, UnaryDP;

def MOVCCi : AI1<0b1101, (outs GPR:$dst),
                        (ins GPR:$false, so_imm:$true), DPFrm,
                "mov", " $dst, $true",
   [/*(set GPR:$dst, (ARMcmov GPR:$false, so_imm:$true, imm:$cc, CCR:$ccr))*/]>,
                RegConstraint<"$false = $dst">, UnaryDP;


//===----------------------------------------------------------------------===//
// TLS Instructions
//

// __aeabi_read_tp preserves the registers r1-r3.
let isCall = 1,
  Defs = [R0, R12, LR, CPSR] in {
  def TPsoft : ABXI<0b1011, (outs), (ins),
               "bl __aeabi_read_tp",
               [(set R0, ARMthread_pointer)]>;
}

//===----------------------------------------------------------------------===//
// SJLJ Exception handling intrinsics
//   eh_sjlj_setjmp() is a three instruction sequence to store the return 
//   address and save #0 in R0 for the non-longjmp case.
//   Since by its nature we may be coming from some other function to get
//   here, and we're using the stack frame for the containing function to
//   save/restore registers, we can't keep anything live in regs across
//   the eh_sjlj_setjmp(), else it will almost certainly have been tromped upon
//   when we get here from a longjmp(). We force everthing out of registers
//   except for our own input by listing the relevant registers in Defs. By
//   doing so, we also cause the prologue/epilogue code to actively preserve
//   all of the callee-saved resgisters, which is exactly what we want.
let Defs = 
  [ R0, R1, R3, R4, R5, R6, R7, R8, R9, R10, R11, R12, LR,
    D0, D2, D3, D4, D5, D6, D7, D8, D9, D10, D11, D12, D13, D14, D15 ] in {
  def Int_eh_sjlj_setjmp : XI<(outs), (ins GPR:$src),
                               AddrModeNone, SizeSpecial, IndexModeNone, Pseudo,
                               "add r0, pc, #4\n\t"
                               "str r0, [$src, #+4]\n\t"
                               "mov r0, #0 @ eh_setjmp", "",
                               [(set R0, (ARMeh_sjlj_setjmp GPR:$src))]>;
}

//===----------------------------------------------------------------------===//
// Non-Instruction Patterns
//

// ConstantPool, GlobalAddress, and JumpTable
def : ARMPat<(ARMWrapper  tglobaladdr :$dst), (LEApcrel tglobaladdr :$dst)>;
def : ARMPat<(ARMWrapper  tconstpool  :$dst), (LEApcrel tconstpool  :$dst)>;
def : ARMPat<(ARMWrapperJT tjumptable:$dst, imm:$id),
             (LEApcrelJT tjumptable:$dst, imm:$id)>;

// Large immediate handling.

// Two piece so_imms.
let isReMaterializable = 1 in
def MOVi2pieces : AI1x2<(outs GPR:$dst), (ins so_imm2part:$src), Pseudo,
                         "mov", " $dst, $src",
                         [(set GPR:$dst, so_imm2part:$src)]>;

def : ARMPat<(or GPR:$LHS, so_imm2part:$RHS),
              (ORRri (ORRri GPR:$LHS, (so_imm2part_1 imm:$RHS)),
                     (so_imm2part_2 imm:$RHS))>;
def : ARMPat<(xor GPR:$LHS, so_imm2part:$RHS),
              (EORri (EORri GPR:$LHS, (so_imm2part_1 imm:$RHS)),
                     (so_imm2part_2 imm:$RHS))>;

// TODO: add,sub,and, 3-instr forms?


// Direct calls
def : ARMPat<(ARMcall texternalsym:$func), (BL texternalsym:$func)>,
      Requires<[IsNotDarwin]>;
def : ARMPat<(ARMcall texternalsym:$func), (BLr9 texternalsym:$func)>,
      Requires<[IsDarwin]>;

// zextload i1 -> zextload i8
def : ARMPat<(zextloadi1 addrmode2:$addr),  (LDRB addrmode2:$addr)>;

// extload -> zextload
def : ARMPat<(extloadi1  addrmode2:$addr),  (LDRB addrmode2:$addr)>;
def : ARMPat<(extloadi8  addrmode2:$addr),  (LDRB addrmode2:$addr)>;
def : ARMPat<(extloadi16 addrmode3:$addr),  (LDRH addrmode3:$addr)>;

def : ARMPat<(extloadi8  addrmodepc:$addr), (PICLDRB addrmodepc:$addr)>;
def : ARMPat<(extloadi16 addrmodepc:$addr), (PICLDRH addrmodepc:$addr)>;

// smul* and smla*
def : ARMV5TEPat<(mul (sra (shl GPR:$a, (i32 16)), (i32 16)),
                      (sra (shl GPR:$b, (i32 16)), (i32 16))),
                 (SMULBB GPR:$a, GPR:$b)>;
def : ARMV5TEPat<(mul sext_16_node:$a, sext_16_node:$b),
                 (SMULBB GPR:$a, GPR:$b)>;
def : ARMV5TEPat<(mul (sra (shl GPR:$a, (i32 16)), (i32 16)),
                      (sra GPR:$b, (i32 16))),
                 (SMULBT GPR:$a, GPR:$b)>;
def : ARMV5TEPat<(mul sext_16_node:$a, (sra GPR:$b, (i32 16))),
                 (SMULBT GPR:$a, GPR:$b)>;
def : ARMV5TEPat<(mul (sra GPR:$a, (i32 16)),
                      (sra (shl GPR:$b, (i32 16)), (i32 16))),
                 (SMULTB GPR:$a, GPR:$b)>;
def : ARMV5TEPat<(mul (sra GPR:$a, (i32 16)), sext_16_node:$b),
                (SMULTB GPR:$a, GPR:$b)>;
def : ARMV5TEPat<(sra (mul GPR:$a, (sra (shl GPR:$b, (i32 16)), (i32 16))),
                      (i32 16)),
                 (SMULWB GPR:$a, GPR:$b)>;
def : ARMV5TEPat<(sra (mul GPR:$a, sext_16_node:$b), (i32 16)),
                 (SMULWB GPR:$a, GPR:$b)>;

def : ARMV5TEPat<(add GPR:$acc,
                      (mul (sra (shl GPR:$a, (i32 16)), (i32 16)),
                           (sra (shl GPR:$b, (i32 16)), (i32 16)))),
                 (SMLABB GPR:$a, GPR:$b, GPR:$acc)>;
def : ARMV5TEPat<(add GPR:$acc,
                      (mul sext_16_node:$a, sext_16_node:$b)),
                 (SMLABB GPR:$a, GPR:$b, GPR:$acc)>;
def : ARMV5TEPat<(add GPR:$acc,
                      (mul (sra (shl GPR:$a, (i32 16)), (i32 16)),
                           (sra GPR:$b, (i32 16)))),
                 (SMLABT GPR:$a, GPR:$b, GPR:$acc)>;
def : ARMV5TEPat<(add GPR:$acc,
                      (mul sext_16_node:$a, (sra GPR:$b, (i32 16)))),
                 (SMLABT GPR:$a, GPR:$b, GPR:$acc)>;
def : ARMV5TEPat<(add GPR:$acc,
                      (mul (sra GPR:$a, (i32 16)),
                           (sra (shl GPR:$b, (i32 16)), (i32 16)))),
                 (SMLATB GPR:$a, GPR:$b, GPR:$acc)>;
def : ARMV5TEPat<(add GPR:$acc,
                      (mul (sra GPR:$a, (i32 16)), sext_16_node:$b)),
                 (SMLATB GPR:$a, GPR:$b, GPR:$acc)>;
def : ARMV5TEPat<(add GPR:$acc,
                      (sra (mul GPR:$a, (sra (shl GPR:$b, (i32 16)), (i32 16))),
                           (i32 16))),
                 (SMLAWB GPR:$a, GPR:$b, GPR:$acc)>;
def : ARMV5TEPat<(add GPR:$acc,
                      (sra (mul GPR:$a, sext_16_node:$b), (i32 16))),
                 (SMLAWB GPR:$a, GPR:$b, GPR:$acc)>;

//===----------------------------------------------------------------------===//
// Thumb Support
//

include "ARMInstrThumb.td"

//===----------------------------------------------------------------------===//
// Thumb2 Support
//

include "ARMInstrThumb2.td"

//===----------------------------------------------------------------------===//
// Floating Point Support
//

include "ARMInstrVFP.td"

//===----------------------------------------------------------------------===//
// Advanced SIMD (NEON) Support
//

include "ARMInstrNEON.td"