Target/RISCV/RISCVInstrInfoD.td

//===-- RISCVInstrInfoD.td - RISC-V 'D' instructions -------*- tablegen -*-===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This file describes the RISC-V instructions from the standard 'D',
// Double-Precision Floating-Point instruction set extension.
//
//===----------------------------------------------------------------------===//

//===----------------------------------------------------------------------===//
// RISC-V specific DAG Nodes.
//===----------------------------------------------------------------------===//

def SDT_RISCVBuildPairF64 : SDTypeProfile<1, 2, [SDTCisVT<0, f64>,
                                                 SDTCisVT<1, i32>,
                                                 SDTCisSameAs<1, 2>]>;
def SDT_RISCVSplitF64     : SDTypeProfile<2, 1, [SDTCisVT<0, i32>,
                                                 SDTCisVT<1, i32>,
                                                 SDTCisVT<2, f64>]>;

def RISCVBuildPairF64 : SDNode<"RISCVISD::BuildPairF64", SDT_RISCVBuildPairF64>;
def RISCVSplitF64     : SDNode<"RISCVISD::SplitF64", SDT_RISCVSplitF64>;

//===----------------------------------------------------------------------===//
// Instruction Class Templates
//===----------------------------------------------------------------------===//

let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in
class FPFMAD_rrr_frm<RISCVOpcode opcode, string opcodestr>
    : RVInstR4<0b01, opcode, (outs FPR64:$rd),
               (ins FPR64:$rs1, FPR64:$rs2, FPR64:$rs3, frmarg:$funct3),
                opcodestr, "$rd, $rs1, $rs2, $rs3, $funct3">;

class FPFMADDynFrmAlias<FPFMAD_rrr_frm Inst, string OpcodeStr>
    : InstAlias<OpcodeStr#" $rd, $rs1, $rs2, $rs3",
                (Inst FPR64:$rd, FPR64:$rs1, FPR64:$rs2, FPR64:$rs3, 0b111)>;

let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in
class FPALUD_rr<bits<7> funct7, bits<3> funct3, string opcodestr>
    : RVInstR<funct7, funct3, OPC_OP_FP, (outs FPR64:$rd),
              (ins FPR64:$rs1, FPR64:$rs2), opcodestr, "$rd, $rs1, $rs2">,
      Sched<[WriteFALU64, ReadFALU64, ReadFALU64]>;

let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in
class FPALUD_rr_frm<bits<7> funct7, string opcodestr>
    : RVInstRFrm<funct7, OPC_OP_FP, (outs FPR64:$rd),
                (ins FPR64:$rs1, FPR64:$rs2, frmarg:$funct3), opcodestr,
                 "$rd, $rs1, $rs2, $funct3">,
      Sched<[WriteFALU64, ReadFALU64, ReadFALU64]>;

class FPALUDDynFrmAlias<FPALUD_rr_frm Inst, string OpcodeStr>
    : InstAlias<OpcodeStr#" $rd, $rs1, $rs2",
                (Inst FPR64:$rd, FPR64:$rs1, FPR64:$rs2, 0b111)>;

let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in
class FPCmpD_rr<bits<3> funct3, string opcodestr>
    : RVInstR<0b1010001, funct3, OPC_OP_FP, (outs GPR:$rd),
              (ins FPR64:$rs1, FPR64:$rs2), opcodestr, "$rd, $rs1, $rs2">,
      Sched<[WriteFCmp64, ReadFCmp64, ReadFCmp64]>;

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

let Predicates = [HasStdExtD] in {

let hasSideEffects = 0, mayLoad = 1, mayStore = 0 in
def FLD : RVInstI<0b011, OPC_LOAD_FP, (outs FPR64:$rd),
                  (ins GPR:$rs1, simm12:$imm12),
                  "fld", "$rd, ${imm12}(${rs1})">,
          Sched<[WriteFLD64, ReadMemBase]>;

// Operands for stores are in the order srcreg, base, offset rather than
// reflecting the order these fields are specified in the instruction
// encoding.
let hasSideEffects = 0, mayLoad = 0, mayStore = 1 in
def FSD : RVInstS<0b011, OPC_STORE_FP, (outs),
                  (ins FPR64:$rs2, GPR:$rs1, simm12:$imm12),
                   "fsd", "$rs2, ${imm12}(${rs1})">,
          Sched<[WriteFST64, ReadStoreData, ReadMemBase]>;

def FMADD_D  : FPFMAD_rrr_frm<OPC_MADD, "fmadd.d">,
               Sched<[WriteFMulAdd64, ReadFMulAdd64, ReadFMulAdd64, ReadFMulAdd64]>;
def          : FPFMADDynFrmAlias<FMADD_D, "fmadd.d">;
def FMSUB_D  : FPFMAD_rrr_frm<OPC_MSUB, "fmsub.d">,
               Sched<[WriteFMulSub64, ReadFMulSub64, ReadFMulSub64, ReadFMulSub64]>;
def          : FPFMADDynFrmAlias<FMSUB_D, "fmsub.d">;
def FNMSUB_D : FPFMAD_rrr_frm<OPC_NMSUB, "fnmsub.d">,
               Sched<[WriteFMulSub64, ReadFMulSub64, ReadFMulSub64, ReadFMulSub64]>;
def          : FPFMADDynFrmAlias<FNMSUB_D, "fnmsub.d">;
def FNMADD_D : FPFMAD_rrr_frm<OPC_NMADD, "fnmadd.d">,
               Sched<[WriteFMulAdd64, ReadFMulAdd64, ReadFMulAdd64, ReadFMulAdd64]>;
def          : FPFMADDynFrmAlias<FNMADD_D, "fnmadd.d">;

def FADD_D : FPALUD_rr_frm<0b0000001, "fadd.d">;
def        : FPALUDDynFrmAlias<FADD_D, "fadd.d">;
def FSUB_D : FPALUD_rr_frm<0b0000101, "fsub.d">;
def        : FPALUDDynFrmAlias<FSUB_D, "fsub.d">;
def FMUL_D : FPALUD_rr_frm<0b0001001, "fmul.d">;
def        : FPALUDDynFrmAlias<FMUL_D, "fmul.d">;
def FDIV_D : FPALUD_rr_frm<0b0001101, "fdiv.d">;
def        : FPALUDDynFrmAlias<FDIV_D, "fdiv.d">;

def FSQRT_D : FPUnaryOp_r_frm<0b0101101, FPR64, FPR64, "fsqrt.d">,
              Sched<[WriteFSqrt32, ReadFSqrt32]> {
  let rs2 = 0b00000;
}
def         : FPUnaryOpDynFrmAlias<FSQRT_D, "fsqrt.d", FPR64, FPR64>;

def FSGNJ_D  : FPALUD_rr<0b0010001, 0b000, "fsgnj.d">;
def FSGNJN_D : FPALUD_rr<0b0010001, 0b001, "fsgnjn.d">;
def FSGNJX_D : FPALUD_rr<0b0010001, 0b010, "fsgnjx.d">;
def FMIN_D   : FPALUD_rr<0b0010101, 0b000, "fmin.d">;
def FMAX_D   : FPALUD_rr<0b0010101, 0b001, "fmax.d">;

def FCVT_S_D : FPUnaryOp_r_frm<0b0100000, FPR32, FPR64, "fcvt.s.d">,
               Sched<[WriteFCvtF64ToF32, ReadFCvtF64ToF32]> {
  let rs2 = 0b00001;
}
def          : FPUnaryOpDynFrmAlias<FCVT_S_D, "fcvt.s.d", FPR32, FPR64>;

def FCVT_D_S : FPUnaryOp_r<0b0100001, 0b000, FPR64, FPR32, "fcvt.d.s">,
               Sched<[WriteFCvtF32ToF64, ReadFCvtF32ToF64]> {
  let rs2 = 0b00000;
}

def FEQ_D : FPCmpD_rr<0b010, "feq.d">;
def FLT_D : FPCmpD_rr<0b001, "flt.d">;
def FLE_D : FPCmpD_rr<0b000, "fle.d">;

def FCLASS_D : FPUnaryOp_r<0b1110001, 0b001, GPR, FPR64, "fclass.d">,
               Sched<[WriteFClass64, ReadFClass64]> {
  let rs2 = 0b00000;
}

def FCVT_W_D : FPUnaryOp_r_frm<0b1100001, GPR, FPR64, "fcvt.w.d">,
               Sched<[WriteFCvtF64ToI32, ReadFCvtF64ToI32]> {
  let rs2 = 0b00000;
}
def          : FPUnaryOpDynFrmAlias<FCVT_W_D, "fcvt.w.d", GPR, FPR64>;

def FCVT_WU_D : FPUnaryOp_r_frm<0b1100001, GPR, FPR64, "fcvt.wu.d">,
                Sched<[WriteFCvtF64ToI32, ReadFCvtF64ToI32]> {
  let rs2 = 0b00001;
}
def           : FPUnaryOpDynFrmAlias<FCVT_WU_D, "fcvt.wu.d", GPR, FPR64>;

def FCVT_D_W : FPUnaryOp_r<0b1101001, 0b000, FPR64, GPR, "fcvt.d.w">,
               Sched<[WriteFCvtI32ToF64, ReadFCvtI32ToF64]> {
  let rs2 = 0b00000;
}

def FCVT_D_WU : FPUnaryOp_r<0b1101001, 0b000, FPR64, GPR, "fcvt.d.wu">,
                Sched<[WriteFCvtI32ToF64, ReadFCvtI32ToF64]> {
  let rs2 = 0b00001;
}
} // Predicates = [HasStdExtD]

let Predicates = [HasStdExtD, IsRV64] in {
def FCVT_L_D : FPUnaryOp_r_frm<0b1100001, GPR, FPR64, "fcvt.l.d">,
               Sched<[WriteFCvtF64ToI64, ReadFCvtF64ToI64]> {
  let rs2 = 0b00010;
}
def          : FPUnaryOpDynFrmAlias<FCVT_L_D, "fcvt.l.d", GPR, FPR64>;

def FCVT_LU_D : FPUnaryOp_r_frm<0b1100001, GPR, FPR64, "fcvt.lu.d">,
                Sched<[WriteFCvtF64ToI64, ReadFCvtF64ToI64]> {
  let rs2 = 0b00011;
}
def           : FPUnaryOpDynFrmAlias<FCVT_LU_D, "fcvt.lu.d", GPR, FPR64>;

def FMV_X_D : FPUnaryOp_r<0b1110001, 0b000, GPR, FPR64, "fmv.x.d">,
              Sched<[WriteFMovF64ToI64, ReadFMovF64ToI64]> {
  let rs2 = 0b00000;
}

def FCVT_D_L : FPUnaryOp_r_frm<0b1101001, FPR64, GPR, "fcvt.d.l">,
               Sched<[WriteFCvtI64ToF64, ReadFCvtI64ToF64]> {
  let rs2 = 0b00010;
}
def          : FPUnaryOpDynFrmAlias<FCVT_D_L, "fcvt.d.l", FPR64, GPR>;

def FCVT_D_LU : FPUnaryOp_r_frm<0b1101001, FPR64, GPR, "fcvt.d.lu">,
                Sched<[WriteFCvtI64ToF64, ReadFCvtI64ToF64]> {
  let rs2 = 0b00011;
}
def           : FPUnaryOpDynFrmAlias<FCVT_D_LU, "fcvt.d.lu", FPR64, GPR>;

def FMV_D_X : FPUnaryOp_r<0b1111001, 0b000, FPR64, GPR, "fmv.d.x">,
              Sched<[WriteFMovI64ToF64, ReadFMovI64ToF64]> {
  let rs2 = 0b00000;
}
} // Predicates = [HasStdExtD, IsRV64]

//===----------------------------------------------------------------------===//
// Assembler Pseudo Instructions (User-Level ISA, Version 2.2, Chapter 20)
//===----------------------------------------------------------------------===//

let Predicates = [HasStdExtD] in {
def : InstAlias<"fld $rd, (${rs1})",  (FLD FPR64:$rd,  GPR:$rs1, 0), 0>;
def : InstAlias<"fsd $rs2, (${rs1})", (FSD FPR64:$rs2, GPR:$rs1, 0), 0>;

def : InstAlias<"fmv.d $rd, $rs",  (FSGNJ_D  FPR64:$rd, FPR64:$rs, FPR64:$rs)>;
def : InstAlias<"fabs.d $rd, $rs", (FSGNJX_D FPR64:$rd, FPR64:$rs, FPR64:$rs)>;
def : InstAlias<"fneg.d $rd, $rs", (FSGNJN_D FPR64:$rd, FPR64:$rs, FPR64:$rs)>;

// fgt.d/fge.d are recognised by the GNU assembler but the canonical
// flt.d/fle.d forms will always be printed. Therefore, set a zero weight.
def : InstAlias<"fgt.d $rd, $rs, $rt",
                (FLT_D GPR:$rd, FPR64:$rt, FPR64:$rs), 0>;
def : InstAlias<"fge.d $rd, $rs, $rt",
                (FLE_D GPR:$rd, FPR64:$rt, FPR64:$rs), 0>;

def PseudoFLD  : PseudoFloatLoad<"fld", FPR64>;
def PseudoFSD  : PseudoStore<"fsd", FPR64>;
} // Predicates = [HasStdExtD]

//===----------------------------------------------------------------------===//
// Pseudo-instructions and codegen patterns
//===----------------------------------------------------------------------===//

class PatFpr64Fpr64<SDPatternOperator OpNode, RVInstR Inst>
    : Pat<(OpNode FPR64:$rs1, FPR64:$rs2), (Inst $rs1, $rs2)>;

class PatFpr64Fpr64DynFrm<SDPatternOperator OpNode, RVInstRFrm Inst>
    : Pat<(OpNode FPR64:$rs1, FPR64:$rs2), (Inst $rs1, $rs2, 0b111)>;

let Predicates = [HasStdExtD] in {

/// Float conversion operations

// f64 -> f32, f32 -> f64
def : Pat<(fpround FPR64:$rs1), (FCVT_S_D FPR64:$rs1, 0b111)>;
def : Pat<(fpextend FPR32:$rs1), (FCVT_D_S FPR32:$rs1)>;

// [u]int<->double conversion patterns must be gated on IsRV32 or IsRV64, so
// are defined later.

/// Float arithmetic operations

def : PatFpr64Fpr64DynFrm<fadd, FADD_D>;
def : PatFpr64Fpr64DynFrm<fsub, FSUB_D>;
def : PatFpr64Fpr64DynFrm<fmul, FMUL_D>;
def : PatFpr64Fpr64DynFrm<fdiv, FDIV_D>;

def : Pat<(fsqrt FPR64:$rs1), (FSQRT_D FPR64:$rs1, 0b111)>;

def : Pat<(fneg FPR64:$rs1), (FSGNJN_D $rs1, $rs1)>;
def : Pat<(fabs FPR64:$rs1), (FSGNJX_D $rs1, $rs1)>;

def : PatFpr64Fpr64<fcopysign, FSGNJ_D>;
def : Pat<(fcopysign FPR64:$rs1, (fneg FPR64:$rs2)), (FSGNJN_D $rs1, $rs2)>;
def : Pat<(fcopysign FPR64:$rs1, FPR32:$rs2), (FSGNJ_D $rs1, (FCVT_D_S $rs2))>;
def : Pat<(fcopysign FPR32:$rs1, FPR64:$rs2), (FSGNJ_S $rs1, (FCVT_S_D $rs2,
                                                              0b111))>;

// fmadd: rs1 * rs2 + rs3
def : Pat<(fma FPR64:$rs1, FPR64:$rs2, FPR64:$rs3),
          (FMADD_D $rs1, $rs2, $rs3, 0b111)>;

// fmsub: rs1 * rs2 - rs3
def : Pat<(fma FPR64:$rs1, FPR64:$rs2, (fneg FPR64:$rs3)),
          (FMSUB_D FPR64:$rs1, FPR64:$rs2, FPR64:$rs3, 0b111)>;

// fnmsub: -rs1 * rs2 + rs3
def : Pat<(fma (fneg FPR64:$rs1), FPR64:$rs2, FPR64:$rs3),
          (FNMSUB_D FPR64:$rs1, FPR64:$rs2, FPR64:$rs3, 0b111)>;

// fnmadd: -rs1 * rs2 - rs3
def : Pat<(fma (fneg FPR64:$rs1), FPR64:$rs2, (fneg FPR64:$rs3)),
          (FNMADD_D FPR64:$rs1, FPR64:$rs2, FPR64:$rs3, 0b111)>;

// The RISC-V 2.2 user-level ISA spec defines fmin and fmax as returning the
// canonical NaN when giving a signaling NaN. This doesn't match the LLVM
// behaviour (see https://bugs.llvm.org/show_bug.cgi?id=27363). However, the
// draft 2.3 ISA spec changes the definition of fmin and fmax in a way that
// matches LLVM's fminnum and fmaxnum
// <https://github.com/riscv/riscv-isa-manual/commit/cd20cee7efd9bac7c5aa127ec3b451749d2b3cce>.
def : PatFpr64Fpr64<fminnum, FMIN_D>;
def : PatFpr64Fpr64<fmaxnum, FMAX_D>;

/// Setcc

def : PatFpr64Fpr64<seteq, FEQ_D>;
def : PatFpr64Fpr64<setoeq, FEQ_D>;
def : PatFpr64Fpr64<setlt, FLT_D>;
def : PatFpr64Fpr64<setolt, FLT_D>;
def : PatFpr64Fpr64<setle, FLE_D>;
def : PatFpr64Fpr64<setole, FLE_D>;

// Define pattern expansions for setcc operations which aren't directly
// handled by a RISC-V instruction and aren't expanded in the SelectionDAG
// Legalizer.

def : Pat<(seto FPR64:$rs1, FPR64:$rs2),
          (AND (FEQ_D FPR64:$rs1, FPR64:$rs1),
               (FEQ_D FPR64:$rs2, FPR64:$rs2))>;

def : Pat<(setuo FPR64:$rs1, FPR64:$rs2),
          (SLTIU (AND (FEQ_D FPR64:$rs1, FPR64:$rs1),
                      (FEQ_D FPR64:$rs2, FPR64:$rs2)),
                 1)>;

def Select_FPR64_Using_CC_GPR : SelectCC_rrirr<FPR64, GPR>;

/// Loads

defm : LdPat<load, FLD>;

/// Stores

defm : StPat<store, FSD, FPR64>;

/// Pseudo-instructions needed for the soft-float ABI with RV32D

// Moves two GPRs to an FPR.
let usesCustomInserter = 1 in
def BuildPairF64Pseudo
    : Pseudo<(outs FPR64:$dst), (ins GPR:$src1, GPR:$src2),
             [(set FPR64:$dst, (RISCVBuildPairF64 GPR:$src1, GPR:$src2))]>;

// Moves an FPR to two GPRs.
let usesCustomInserter = 1 in
def SplitF64Pseudo
    : Pseudo<(outs GPR:$dst1, GPR:$dst2), (ins FPR64:$src),
             [(set GPR:$dst1, GPR:$dst2, (RISCVSplitF64 FPR64:$src))]>;

} // Predicates = [HasStdExtD]

let Predicates = [HasStdExtD, IsRV32] in {
// double->[u]int. Round-to-zero must be used.
def : Pat<(fp_to_sint FPR64:$rs1), (FCVT_W_D FPR64:$rs1, 0b001)>;
def : Pat<(fp_to_uint FPR64:$rs1), (FCVT_WU_D FPR64:$rs1, 0b001)>;

// [u]int->double.
def : Pat<(sint_to_fp GPR:$rs1), (FCVT_D_W GPR:$rs1)>;
def : Pat<(uint_to_fp GPR:$rs1), (FCVT_D_WU GPR:$rs1)>;
} // Predicates = [HasStdExtD, IsRV32]

let Predicates = [HasStdExtD, IsRV64] in {
def : Pat<(bitconvert GPR:$rs1), (FMV_D_X GPR:$rs1)>;
def : Pat<(bitconvert FPR64:$rs1), (FMV_X_D FPR64:$rs1)>;

// FP->[u]int32 is mostly handled by the FP->[u]int64 patterns. This is safe
// because fpto[u|s]i produce poison if the value can't fit into the target.
// We match the single case below because fcvt.wu.d sign-extends its result so
// is cheaper than fcvt.lu.d+sext.w.
def : Pat<(sext_inreg (zexti32 (fp_to_uint FPR64:$rs1)), i32),
          (FCVT_WU_D $rs1, 0b001)>;

// [u]int32->fp
def : Pat<(sint_to_fp (sext_inreg GPR:$rs1, i32)), (FCVT_D_W $rs1)>;
def : Pat<(uint_to_fp (zexti32 GPR:$rs1)), (FCVT_D_WU $rs1)>;

def : Pat<(fp_to_sint FPR64:$rs1), (FCVT_L_D FPR64:$rs1, 0b001)>;
def : Pat<(fp_to_uint FPR64:$rs1), (FCVT_LU_D FPR64:$rs1, 0b001)>;

// [u]int64->fp. Match GCC and default to using dynamic rounding mode.
def : Pat<(sint_to_fp GPR:$rs1), (FCVT_D_L GPR:$rs1, 0b111)>;
def : Pat<(uint_to_fp GPR:$rs1), (FCVT_D_LU GPR:$rs1, 0b111)>;
} // Predicates = [HasStdExtD, IsRV64]