diff options
Diffstat (limited to 'lib/Target/X86/X86InstrFPStack.td')
| -rw-r--r-- | lib/Target/X86/X86InstrFPStack.td | 597 |
1 files changed, 597 insertions, 0 deletions
diff --git a/lib/Target/X86/X86InstrFPStack.td b/lib/Target/X86/X86InstrFPStack.td new file mode 100644 index 0000000..bc7def4 --- /dev/null +++ b/lib/Target/X86/X86InstrFPStack.td @@ -0,0 +1,597 @@ +//==- X86InstrFPStack.td - Describe the X86 Instruction Set --*- tablegen -*-=// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file describes the X86 x87 FPU instruction set, defining the +// instructions, and properties of the instructions which are needed for code +// generation, machine code emission, and analysis. +// +//===----------------------------------------------------------------------===// + +//===----------------------------------------------------------------------===// +// FPStack specific DAG Nodes. +//===----------------------------------------------------------------------===// + +def SDTX86FpGet2 : SDTypeProfile<2, 0, [SDTCisVT<0, f80>, + SDTCisVT<1, f80>]>; +def SDTX86Fld : SDTypeProfile<1, 2, [SDTCisFP<0>, + SDTCisPtrTy<1>, + SDTCisVT<2, OtherVT>]>; +def SDTX86Fst : SDTypeProfile<0, 3, [SDTCisFP<0>, + SDTCisPtrTy<1>, + SDTCisVT<2, OtherVT>]>; +def SDTX86Fild : SDTypeProfile<1, 2, [SDTCisFP<0>, SDTCisPtrTy<1>, + SDTCisVT<2, OtherVT>]>; +def SDTX86FpToIMem : SDTypeProfile<0, 2, [SDTCisFP<0>, SDTCisPtrTy<1>]>; + +def SDTX86CwdStore : SDTypeProfile<0, 1, [SDTCisPtrTy<0>]>; + +def X86fld : SDNode<"X86ISD::FLD", SDTX86Fld, + [SDNPHasChain, SDNPMayLoad]>; +def X86fst : SDNode<"X86ISD::FST", SDTX86Fst, + [SDNPHasChain, SDNPInFlag, SDNPMayStore]>; +def X86fild : SDNode<"X86ISD::FILD", SDTX86Fild, + [SDNPHasChain, SDNPMayLoad]>; +def X86fildflag : SDNode<"X86ISD::FILD_FLAG", SDTX86Fild, + [SDNPHasChain, SDNPOutFlag, SDNPMayLoad]>; +def X86fp_to_i16mem : SDNode<"X86ISD::FP_TO_INT16_IN_MEM", SDTX86FpToIMem, + [SDNPHasChain, SDNPMayStore]>; +def X86fp_to_i32mem : SDNode<"X86ISD::FP_TO_INT32_IN_MEM", SDTX86FpToIMem, + [SDNPHasChain, SDNPMayStore]>; +def X86fp_to_i64mem : SDNode<"X86ISD::FP_TO_INT64_IN_MEM", SDTX86FpToIMem, + [SDNPHasChain, SDNPMayStore]>; +def X86fp_cwd_get16 : SDNode<"X86ISD::FNSTCW16m", SDTX86CwdStore, + [SDNPHasChain, SDNPMayStore, SDNPSideEffect]>; + +//===----------------------------------------------------------------------===// +// FPStack pattern fragments +//===----------------------------------------------------------------------===// + +def fpimm0 : PatLeaf<(fpimm), [{ + return N->isExactlyValue(+0.0); +}]>; + +def fpimmneg0 : PatLeaf<(fpimm), [{ + return N->isExactlyValue(-0.0); +}]>; + +def fpimm1 : PatLeaf<(fpimm), [{ + return N->isExactlyValue(+1.0); +}]>; + +def fpimmneg1 : PatLeaf<(fpimm), [{ + return N->isExactlyValue(-1.0); +}]>; + +// Some 'special' instructions +let usesCustomDAGSchedInserter = 1 in { // Expanded by the scheduler. + def FP32_TO_INT16_IN_MEM : I<0, Pseudo, + (outs), (ins i16mem:$dst, RFP32:$src), + "##FP32_TO_INT16_IN_MEM PSEUDO!", + [(X86fp_to_i16mem RFP32:$src, addr:$dst)]>; + def FP32_TO_INT32_IN_MEM : I<0, Pseudo, + (outs), (ins i32mem:$dst, RFP32:$src), + "##FP32_TO_INT32_IN_MEM PSEUDO!", + [(X86fp_to_i32mem RFP32:$src, addr:$dst)]>; + def FP32_TO_INT64_IN_MEM : I<0, Pseudo, + (outs), (ins i64mem:$dst, RFP32:$src), + "##FP32_TO_INT64_IN_MEM PSEUDO!", + [(X86fp_to_i64mem RFP32:$src, addr:$dst)]>; + def FP64_TO_INT16_IN_MEM : I<0, Pseudo, + (outs), (ins i16mem:$dst, RFP64:$src), + "##FP64_TO_INT16_IN_MEM PSEUDO!", + [(X86fp_to_i16mem RFP64:$src, addr:$dst)]>; + def FP64_TO_INT32_IN_MEM : I<0, Pseudo, + (outs), (ins i32mem:$dst, RFP64:$src), + "##FP64_TO_INT32_IN_MEM PSEUDO!", + [(X86fp_to_i32mem RFP64:$src, addr:$dst)]>; + def FP64_TO_INT64_IN_MEM : I<0, Pseudo, + (outs), (ins i64mem:$dst, RFP64:$src), + "##FP64_TO_INT64_IN_MEM PSEUDO!", + [(X86fp_to_i64mem RFP64:$src, addr:$dst)]>; + def FP80_TO_INT16_IN_MEM : I<0, Pseudo, + (outs), (ins i16mem:$dst, RFP80:$src), + "##FP80_TO_INT16_IN_MEM PSEUDO!", + [(X86fp_to_i16mem RFP80:$src, addr:$dst)]>; + def FP80_TO_INT32_IN_MEM : I<0, Pseudo, + (outs), (ins i32mem:$dst, RFP80:$src), + "##FP80_TO_INT32_IN_MEM PSEUDO!", + [(X86fp_to_i32mem RFP80:$src, addr:$dst)]>; + def FP80_TO_INT64_IN_MEM : I<0, Pseudo, + (outs), (ins i64mem:$dst, RFP80:$src), + "##FP80_TO_INT64_IN_MEM PSEUDO!", + [(X86fp_to_i64mem RFP80:$src, addr:$dst)]>; +} + +let isTerminator = 1 in + let Defs = [FP0, FP1, FP2, FP3, FP4, FP5, FP6] in + def FP_REG_KILL : I<0, Pseudo, (outs), (ins), "##FP_REG_KILL", []>; + +// All FP Stack operations are represented with four instructions here. The +// first three instructions, generated by the instruction selector, use "RFP32" +// "RFP64" or "RFP80" registers: traditional register files to reference 32-bit, +// 64-bit or 80-bit floating point values. These sizes apply to the values, +// not the registers, which are always 80 bits; RFP32, RFP64 and RFP80 can be +// copied to each other without losing information. These instructions are all +// pseudo instructions and use the "_Fp" suffix. +// In some cases there are additional variants with a mixture of different +// register sizes. +// The second instruction is defined with FPI, which is the actual instruction +// emitted by the assembler. These use "RST" registers, although frequently +// the actual register(s) used are implicit. These are always 80 bits. +// The FP stackifier pass converts one to the other after register allocation +// occurs. +// +// Note that the FpI instruction should have instruction selection info (e.g. +// a pattern) and the FPI instruction should have emission info (e.g. opcode +// encoding and asm printing info). + +// Pseudo Instructions for FP stack return values. +def FpGET_ST0_32 : FpI_<(outs RFP32:$dst), (ins), SpecialFP, []>; // FPR = ST(0) +def FpGET_ST0_64 : FpI_<(outs RFP64:$dst), (ins), SpecialFP, []>; // FPR = ST(0) +def FpGET_ST0_80 : FpI_<(outs RFP80:$dst), (ins), SpecialFP, []>; // FPR = ST(0) + +// FpGET_ST1* should only be issued *after* an FpGET_ST0* has been issued when +// there are two values live out on the stack from a call or inlineasm. This +// magic is handled by the stackifier. It is not valid to emit FpGET_ST1* and +// then FpGET_ST0*. In addition, it is invalid for any FP-using operations to +// occur between them. +def FpGET_ST1_32 : FpI_<(outs RFP32:$dst), (ins), SpecialFP, []>; // FPR = ST(1) +def FpGET_ST1_64 : FpI_<(outs RFP64:$dst), (ins), SpecialFP, []>; // FPR = ST(1) +def FpGET_ST1_80 : FpI_<(outs RFP80:$dst), (ins), SpecialFP, []>; // FPR = ST(1) + +let Defs = [ST0] in { +def FpSET_ST0_32 : FpI_<(outs), (ins RFP32:$src), SpecialFP, []>; // ST(0) = FPR +def FpSET_ST0_64 : FpI_<(outs), (ins RFP64:$src), SpecialFP, []>; // ST(0) = FPR +def FpSET_ST0_80 : FpI_<(outs), (ins RFP80:$src), SpecialFP, []>; // ST(0) = FPR +} + +let Defs = [ST1] in { +def FpSET_ST1_32 : FpI_<(outs), (ins RFP32:$src), SpecialFP, []>; // ST(1) = FPR +def FpSET_ST1_64 : FpI_<(outs), (ins RFP64:$src), SpecialFP, []>; // ST(1) = FPR +def FpSET_ST1_80 : FpI_<(outs), (ins RFP80:$src), SpecialFP, []>; // ST(1) = FPR +} + +// FpIf32, FpIf64 - Floating Point Psuedo Instruction template. +// f32 instructions can use SSE1 and are predicated on FPStackf32 == !SSE1. +// f64 instructions can use SSE2 and are predicated on FPStackf64 == !SSE2. +// f80 instructions cannot use SSE and use neither of these. +class FpIf32<dag outs, dag ins, FPFormat fp, list<dag> pattern> : + FpI_<outs, ins, fp, pattern>, Requires<[FPStackf32]>; +class FpIf64<dag outs, dag ins, FPFormat fp, list<dag> pattern> : + FpI_<outs, ins, fp, pattern>, Requires<[FPStackf64]>; + +// Register copies. Just copies, the shortening ones do not truncate. +let neverHasSideEffects = 1 in { + def MOV_Fp3232 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src), SpecialFP, []>; + def MOV_Fp3264 : FpIf32<(outs RFP64:$dst), (ins RFP32:$src), SpecialFP, []>; + def MOV_Fp6432 : FpIf32<(outs RFP32:$dst), (ins RFP64:$src), SpecialFP, []>; + def MOV_Fp6464 : FpIf64<(outs RFP64:$dst), (ins RFP64:$src), SpecialFP, []>; + def MOV_Fp8032 : FpIf32<(outs RFP32:$dst), (ins RFP80:$src), SpecialFP, []>; + def MOV_Fp3280 : FpIf32<(outs RFP80:$dst), (ins RFP32:$src), SpecialFP, []>; + def MOV_Fp8064 : FpIf64<(outs RFP64:$dst), (ins RFP80:$src), SpecialFP, []>; + def MOV_Fp6480 : FpIf64<(outs RFP80:$dst), (ins RFP64:$src), SpecialFP, []>; + def MOV_Fp8080 : FpI_ <(outs RFP80:$dst), (ins RFP80:$src), SpecialFP, []>; +} + +// Factoring for arithmetic. +multiclass FPBinary_rr<SDNode OpNode> { +// Register op register -> register +// These are separated out because they have no reversed form. +def _Fp32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, RFP32:$src2), TwoArgFP, + [(set RFP32:$dst, (OpNode RFP32:$src1, RFP32:$src2))]>; +def _Fp64 : FpIf64<(outs RFP64:$dst), (ins RFP64:$src1, RFP64:$src2), TwoArgFP, + [(set RFP64:$dst, (OpNode RFP64:$src1, RFP64:$src2))]>; +def _Fp80 : FpI_<(outs RFP80:$dst), (ins RFP80:$src1, RFP80:$src2), TwoArgFP, + [(set RFP80:$dst, (OpNode RFP80:$src1, RFP80:$src2))]>; +} +// The FopST0 series are not included here because of the irregularities +// in where the 'r' goes in assembly output. +// These instructions cannot address 80-bit memory. +multiclass FPBinary<SDNode OpNode, Format fp, string asmstring> { +// ST(0) = ST(0) + [mem] +def _Fp32m : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, f32mem:$src2), OneArgFPRW, + [(set RFP32:$dst, + (OpNode RFP32:$src1, (loadf32 addr:$src2)))]>; +def _Fp64m : FpIf64<(outs RFP64:$dst), (ins RFP64:$src1, f64mem:$src2), OneArgFPRW, + [(set RFP64:$dst, + (OpNode RFP64:$src1, (loadf64 addr:$src2)))]>; +def _Fp64m32: FpIf64<(outs RFP64:$dst), (ins RFP64:$src1, f32mem:$src2), OneArgFPRW, + [(set RFP64:$dst, + (OpNode RFP64:$src1, (f64 (extloadf32 addr:$src2))))]>; +def _Fp80m32: FpI_<(outs RFP80:$dst), (ins RFP80:$src1, f32mem:$src2), OneArgFPRW, + [(set RFP80:$dst, + (OpNode RFP80:$src1, (f80 (extloadf32 addr:$src2))))]>; +def _Fp80m64: FpI_<(outs RFP80:$dst), (ins RFP80:$src1, f64mem:$src2), OneArgFPRW, + [(set RFP80:$dst, + (OpNode RFP80:$src1, (f80 (extloadf64 addr:$src2))))]>; +def _F32m : FPI<0xD8, fp, (outs), (ins f32mem:$src), + !strconcat("f", !strconcat(asmstring, "{s}\t$src"))> { let mayLoad = 1; } +def _F64m : FPI<0xDC, fp, (outs), (ins f64mem:$src), + !strconcat("f", !strconcat(asmstring, "{l}\t$src"))> { let mayLoad = 1; } +// ST(0) = ST(0) + [memint] +def _FpI16m32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, i16mem:$src2), OneArgFPRW, + [(set RFP32:$dst, (OpNode RFP32:$src1, + (X86fild addr:$src2, i16)))]>; +def _FpI32m32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, i32mem:$src2), OneArgFPRW, + [(set RFP32:$dst, (OpNode RFP32:$src1, + (X86fild addr:$src2, i32)))]>; +def _FpI16m64 : FpIf64<(outs RFP64:$dst), (ins RFP64:$src1, i16mem:$src2), OneArgFPRW, + [(set RFP64:$dst, (OpNode RFP64:$src1, + (X86fild addr:$src2, i16)))]>; +def _FpI32m64 : FpIf64<(outs RFP64:$dst), (ins RFP64:$src1, i32mem:$src2), OneArgFPRW, + [(set RFP64:$dst, (OpNode RFP64:$src1, + (X86fild addr:$src2, i32)))]>; +def _FpI16m80 : FpI_<(outs RFP80:$dst), (ins RFP80:$src1, i16mem:$src2), OneArgFPRW, + [(set RFP80:$dst, (OpNode RFP80:$src1, + (X86fild addr:$src2, i16)))]>; +def _FpI32m80 : FpI_<(outs RFP80:$dst), (ins RFP80:$src1, i32mem:$src2), OneArgFPRW, + [(set RFP80:$dst, (OpNode RFP80:$src1, + (X86fild addr:$src2, i32)))]>; +def _FI16m : FPI<0xDE, fp, (outs), (ins i16mem:$src), + !strconcat("fi", !strconcat(asmstring, "{s}\t$src"))> { let mayLoad = 1; } +def _FI32m : FPI<0xDA, fp, (outs), (ins i32mem:$src), + !strconcat("fi", !strconcat(asmstring, "{l}\t$src"))> { let mayLoad = 1; } +} + +defm ADD : FPBinary_rr<fadd>; +defm SUB : FPBinary_rr<fsub>; +defm MUL : FPBinary_rr<fmul>; +defm DIV : FPBinary_rr<fdiv>; +defm ADD : FPBinary<fadd, MRM0m, "add">; +defm SUB : FPBinary<fsub, MRM4m, "sub">; +defm SUBR: FPBinary<fsub ,MRM5m, "subr">; +defm MUL : FPBinary<fmul, MRM1m, "mul">; +defm DIV : FPBinary<fdiv, MRM6m, "div">; +defm DIVR: FPBinary<fdiv, MRM7m, "divr">; + +class FPST0rInst<bits<8> o, string asm> + : FPI<o, AddRegFrm, (outs), (ins RST:$op), asm>, D8; +class FPrST0Inst<bits<8> o, string asm> + : FPI<o, AddRegFrm, (outs), (ins RST:$op), asm>, DC; +class FPrST0PInst<bits<8> o, string asm> + : FPI<o, AddRegFrm, (outs), (ins RST:$op), asm>, DE; + +// NOTE: GAS and apparently all other AT&T style assemblers have a broken notion +// of some of the 'reverse' forms of the fsub and fdiv instructions. As such, +// we have to put some 'r's in and take them out of weird places. +def ADD_FST0r : FPST0rInst <0xC0, "fadd\t$op">; +def ADD_FrST0 : FPrST0Inst <0xC0, "fadd\t{%st(0), $op|$op, %ST(0)}">; +def ADD_FPrST0 : FPrST0PInst<0xC0, "faddp\t$op">; +def SUBR_FST0r : FPST0rInst <0xE8, "fsubr\t$op">; +def SUB_FrST0 : FPrST0Inst <0xE8, "fsub{r}\t{%st(0), $op|$op, %ST(0)}">; +def SUB_FPrST0 : FPrST0PInst<0xE8, "fsub{r}p\t$op">; +def SUB_FST0r : FPST0rInst <0xE0, "fsub\t$op">; +def SUBR_FrST0 : FPrST0Inst <0xE0, "fsub{|r}\t{%st(0), $op|$op, %ST(0)}">; +def SUBR_FPrST0 : FPrST0PInst<0xE0, "fsub{|r}p\t$op">; +def MUL_FST0r : FPST0rInst <0xC8, "fmul\t$op">; +def MUL_FrST0 : FPrST0Inst <0xC8, "fmul\t{%st(0), $op|$op, %ST(0)}">; +def MUL_FPrST0 : FPrST0PInst<0xC8, "fmulp\t$op">; +def DIVR_FST0r : FPST0rInst <0xF8, "fdivr\t$op">; +def DIV_FrST0 : FPrST0Inst <0xF8, "fdiv{r}\t{%st(0), $op|$op, %ST(0)}">; +def DIV_FPrST0 : FPrST0PInst<0xF8, "fdiv{r}p\t$op">; +def DIV_FST0r : FPST0rInst <0xF0, "fdiv\t$op">; +def DIVR_FrST0 : FPrST0Inst <0xF0, "fdiv{|r}\t{%st(0), $op|$op, %ST(0)}">; +def DIVR_FPrST0 : FPrST0PInst<0xF0, "fdiv{|r}p\t$op">; + +// Unary operations. +multiclass FPUnary<SDNode OpNode, bits<8> opcode, string asmstring> { +def _Fp32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src), OneArgFPRW, + [(set RFP32:$dst, (OpNode RFP32:$src))]>; +def _Fp64 : FpIf64<(outs RFP64:$dst), (ins RFP64:$src), OneArgFPRW, + [(set RFP64:$dst, (OpNode RFP64:$src))]>; +def _Fp80 : FpI_<(outs RFP80:$dst), (ins RFP80:$src), OneArgFPRW, + [(set RFP80:$dst, (OpNode RFP80:$src))]>; +def _F : FPI<opcode, RawFrm, (outs), (ins), asmstring>, D9; +} + +defm CHS : FPUnary<fneg, 0xE0, "fchs">; +defm ABS : FPUnary<fabs, 0xE1, "fabs">; +defm SQRT: FPUnary<fsqrt,0xFA, "fsqrt">; +defm SIN : FPUnary<fsin, 0xFE, "fsin">; +defm COS : FPUnary<fcos, 0xFF, "fcos">; + +let neverHasSideEffects = 1 in { +def TST_Fp32 : FpIf32<(outs), (ins RFP32:$src), OneArgFP, []>; +def TST_Fp64 : FpIf64<(outs), (ins RFP64:$src), OneArgFP, []>; +def TST_Fp80 : FpI_<(outs), (ins RFP80:$src), OneArgFP, []>; +} +def TST_F : FPI<0xE4, RawFrm, (outs), (ins), "ftst">, D9; + +// Floating point cmovs. +multiclass FPCMov<PatLeaf cc> { + def _Fp32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, RFP32:$src2), + CondMovFP, + [(set RFP32:$dst, (X86cmov RFP32:$src1, RFP32:$src2, + cc, EFLAGS))]>; + def _Fp64 : FpIf64<(outs RFP64:$dst), (ins RFP64:$src1, RFP64:$src2), + CondMovFP, + [(set RFP64:$dst, (X86cmov RFP64:$src1, RFP64:$src2, + cc, EFLAGS))]>; + def _Fp80 : FpI_<(outs RFP80:$dst), (ins RFP80:$src1, RFP80:$src2), + CondMovFP, + [(set RFP80:$dst, (X86cmov RFP80:$src1, RFP80:$src2, + cc, EFLAGS))]>; +} +let Uses = [EFLAGS], isTwoAddress = 1 in { +defm CMOVB : FPCMov<X86_COND_B>; +defm CMOVBE : FPCMov<X86_COND_BE>; +defm CMOVE : FPCMov<X86_COND_E>; +defm CMOVP : FPCMov<X86_COND_P>; +defm CMOVNB : FPCMov<X86_COND_AE>; +defm CMOVNBE: FPCMov<X86_COND_A>; +defm CMOVNE : FPCMov<X86_COND_NE>; +defm CMOVNP : FPCMov<X86_COND_NP>; +} + +// These are not factored because there's no clean way to pass DA/DB. +def CMOVB_F : FPI<0xC0, AddRegFrm, (outs RST:$op), (ins), + "fcmovb\t{$op, %st(0)|%ST(0), $op}">, DA; +def CMOVBE_F : FPI<0xD0, AddRegFrm, (outs RST:$op), (ins), + "fcmovbe\t{$op, %st(0)|%ST(0), $op}">, DA; +def CMOVE_F : FPI<0xC8, AddRegFrm, (outs RST:$op), (ins), + "fcmove\t{$op, %st(0)|%ST(0), $op}">, DA; +def CMOVP_F : FPI<0xD8, AddRegFrm, (outs RST:$op), (ins), + "fcmovu\t {$op, %st(0)|%ST(0), $op}">, DA; +def CMOVNB_F : FPI<0xC0, AddRegFrm, (outs RST:$op), (ins), + "fcmovnb\t{$op, %st(0)|%ST(0), $op}">, DB; +def CMOVNBE_F: FPI<0xD0, AddRegFrm, (outs RST:$op), (ins), + "fcmovnbe\t{$op, %st(0)|%ST(0), $op}">, DB; +def CMOVNE_F : FPI<0xC8, AddRegFrm, (outs RST:$op), (ins), + "fcmovne\t{$op, %st(0)|%ST(0), $op}">, DB; +def CMOVNP_F : FPI<0xD8, AddRegFrm, (outs RST:$op), (ins), + "fcmovnu\t{$op, %st(0)|%ST(0), $op}">, DB; + +// Floating point loads & stores. +let canFoldAsLoad = 1 in { +def LD_Fp32m : FpIf32<(outs RFP32:$dst), (ins f32mem:$src), ZeroArgFP, + [(set RFP32:$dst, (loadf32 addr:$src))]>; +let isReMaterializable = 1, mayHaveSideEffects = 1 in + def LD_Fp64m : FpIf64<(outs RFP64:$dst), (ins f64mem:$src), ZeroArgFP, + [(set RFP64:$dst, (loadf64 addr:$src))]>; +def LD_Fp80m : FpI_<(outs RFP80:$dst), (ins f80mem:$src), ZeroArgFP, + [(set RFP80:$dst, (loadf80 addr:$src))]>; +} +def LD_Fp32m64 : FpIf64<(outs RFP64:$dst), (ins f32mem:$src), ZeroArgFP, + [(set RFP64:$dst, (f64 (extloadf32 addr:$src)))]>; +def LD_Fp64m80 : FpI_<(outs RFP80:$dst), (ins f64mem:$src), ZeroArgFP, + [(set RFP80:$dst, (f80 (extloadf64 addr:$src)))]>; +def LD_Fp32m80 : FpI_<(outs RFP80:$dst), (ins f32mem:$src), ZeroArgFP, + [(set RFP80:$dst, (f80 (extloadf32 addr:$src)))]>; +def ILD_Fp16m32: FpIf32<(outs RFP32:$dst), (ins i16mem:$src), ZeroArgFP, + [(set RFP32:$dst, (X86fild addr:$src, i16))]>; +def ILD_Fp32m32: FpIf32<(outs RFP32:$dst), (ins i32mem:$src), ZeroArgFP, + [(set RFP32:$dst, (X86fild addr:$src, i32))]>; +def ILD_Fp64m32: FpIf32<(outs RFP32:$dst), (ins i64mem:$src), ZeroArgFP, + [(set RFP32:$dst, (X86fild addr:$src, i64))]>; +def ILD_Fp16m64: FpIf64<(outs RFP64:$dst), (ins i16mem:$src), ZeroArgFP, + [(set RFP64:$dst, (X86fild addr:$src, i16))]>; +def ILD_Fp32m64: FpIf64<(outs RFP64:$dst), (ins i32mem:$src), ZeroArgFP, + [(set RFP64:$dst, (X86fild addr:$src, i32))]>; +def ILD_Fp64m64: FpIf64<(outs RFP64:$dst), (ins i64mem:$src), ZeroArgFP, + [(set RFP64:$dst, (X86fild addr:$src, i64))]>; +def ILD_Fp16m80: FpI_<(outs RFP80:$dst), (ins i16mem:$src), ZeroArgFP, + [(set RFP80:$dst, (X86fild addr:$src, i16))]>; +def ILD_Fp32m80: FpI_<(outs RFP80:$dst), (ins i32mem:$src), ZeroArgFP, + [(set RFP80:$dst, (X86fild addr:$src, i32))]>; +def ILD_Fp64m80: FpI_<(outs RFP80:$dst), (ins i64mem:$src), ZeroArgFP, + [(set RFP80:$dst, (X86fild addr:$src, i64))]>; + +def ST_Fp32m : FpIf32<(outs), (ins f32mem:$op, RFP32:$src), OneArgFP, + [(store RFP32:$src, addr:$op)]>; +def ST_Fp64m32 : FpIf64<(outs), (ins f32mem:$op, RFP64:$src), OneArgFP, + [(truncstoref32 RFP64:$src, addr:$op)]>; +def ST_Fp64m : FpIf64<(outs), (ins f64mem:$op, RFP64:$src), OneArgFP, + [(store RFP64:$src, addr:$op)]>; +def ST_Fp80m32 : FpI_<(outs), (ins f32mem:$op, RFP80:$src), OneArgFP, + [(truncstoref32 RFP80:$src, addr:$op)]>; +def ST_Fp80m64 : FpI_<(outs), (ins f64mem:$op, RFP80:$src), OneArgFP, + [(truncstoref64 RFP80:$src, addr:$op)]>; +// FST does not support 80-bit memory target; FSTP must be used. + +let mayStore = 1, neverHasSideEffects = 1 in { +def ST_FpP32m : FpIf32<(outs), (ins f32mem:$op, RFP32:$src), OneArgFP, []>; +def ST_FpP64m32 : FpIf64<(outs), (ins f32mem:$op, RFP64:$src), OneArgFP, []>; +def ST_FpP64m : FpIf64<(outs), (ins f64mem:$op, RFP64:$src), OneArgFP, []>; +def ST_FpP80m32 : FpI_<(outs), (ins f32mem:$op, RFP80:$src), OneArgFP, []>; +def ST_FpP80m64 : FpI_<(outs), (ins f64mem:$op, RFP80:$src), OneArgFP, []>; +} +def ST_FpP80m : FpI_<(outs), (ins f80mem:$op, RFP80:$src), OneArgFP, + [(store RFP80:$src, addr:$op)]>; +let mayStore = 1, neverHasSideEffects = 1 in { +def IST_Fp16m32 : FpIf32<(outs), (ins i16mem:$op, RFP32:$src), OneArgFP, []>; +def IST_Fp32m32 : FpIf32<(outs), (ins i32mem:$op, RFP32:$src), OneArgFP, []>; +def IST_Fp64m32 : FpIf32<(outs), (ins i64mem:$op, RFP32:$src), OneArgFP, []>; +def IST_Fp16m64 : FpIf64<(outs), (ins i16mem:$op, RFP64:$src), OneArgFP, []>; +def IST_Fp32m64 : FpIf64<(outs), (ins i32mem:$op, RFP64:$src), OneArgFP, []>; +def IST_Fp64m64 : FpIf64<(outs), (ins i64mem:$op, RFP64:$src), OneArgFP, []>; +def IST_Fp16m80 : FpI_<(outs), (ins i16mem:$op, RFP80:$src), OneArgFP, []>; +def IST_Fp32m80 : FpI_<(outs), (ins i32mem:$op, RFP80:$src), OneArgFP, []>; +def IST_Fp64m80 : FpI_<(outs), (ins i64mem:$op, RFP80:$src), OneArgFP, []>; +} + +let mayLoad = 1 in { +def LD_F32m : FPI<0xD9, MRM0m, (outs), (ins f32mem:$src), "fld{s}\t$src">; +def LD_F64m : FPI<0xDD, MRM0m, (outs), (ins f64mem:$src), "fld{l}\t$src">; +def LD_F80m : FPI<0xDB, MRM5m, (outs), (ins f80mem:$src), "fld{t}\t$src">; +def ILD_F16m : FPI<0xDF, MRM0m, (outs), (ins i16mem:$src), "fild{s}\t$src">; +def ILD_F32m : FPI<0xDB, MRM0m, (outs), (ins i32mem:$src), "fild{l}\t$src">; +def ILD_F64m : FPI<0xDF, MRM5m, (outs), (ins i64mem:$src), "fild{ll}\t$src">; +} +let mayStore = 1 in { +def ST_F32m : FPI<0xD9, MRM2m, (outs), (ins f32mem:$dst), "fst{s}\t$dst">; +def ST_F64m : FPI<0xDD, MRM2m, (outs), (ins f64mem:$dst), "fst{l}\t$dst">; +def ST_FP32m : FPI<0xD9, MRM3m, (outs), (ins f32mem:$dst), "fstp{s}\t$dst">; +def ST_FP64m : FPI<0xDD, MRM3m, (outs), (ins f64mem:$dst), "fstp{l}\t$dst">; +def ST_FP80m : FPI<0xDB, MRM7m, (outs), (ins f80mem:$dst), "fstp{t}\t$dst">; +def IST_F16m : FPI<0xDF, MRM2m, (outs), (ins i16mem:$dst), "fist{s}\t$dst">; +def IST_F32m : FPI<0xDB, MRM2m, (outs), (ins i32mem:$dst), "fist{l}\t$dst">; +def IST_FP16m : FPI<0xDF, MRM3m, (outs), (ins i16mem:$dst), "fistp{s}\t$dst">; +def IST_FP32m : FPI<0xDB, MRM3m, (outs), (ins i32mem:$dst), "fistp{l}\t$dst">; +def IST_FP64m : FPI<0xDF, MRM7m, (outs), (ins i64mem:$dst), "fistp{ll}\t$dst">; +} + +// FISTTP requires SSE3 even though it's a FPStack op. +def ISTT_Fp16m32 : FpI_<(outs), (ins i16mem:$op, RFP32:$src), OneArgFP, + [(X86fp_to_i16mem RFP32:$src, addr:$op)]>, + Requires<[HasSSE3]>; +def ISTT_Fp32m32 : FpI_<(outs), (ins i32mem:$op, RFP32:$src), OneArgFP, + [(X86fp_to_i32mem RFP32:$src, addr:$op)]>, + Requires<[HasSSE3]>; +def ISTT_Fp64m32 : FpI_<(outs), (ins i64mem:$op, RFP32:$src), OneArgFP, + [(X86fp_to_i64mem RFP32:$src, addr:$op)]>, + Requires<[HasSSE3]>; +def ISTT_Fp16m64 : FpI_<(outs), (ins i16mem:$op, RFP64:$src), OneArgFP, + [(X86fp_to_i16mem RFP64:$src, addr:$op)]>, + Requires<[HasSSE3]>; +def ISTT_Fp32m64 : FpI_<(outs), (ins i32mem:$op, RFP64:$src), OneArgFP, + [(X86fp_to_i32mem RFP64:$src, addr:$op)]>, + Requires<[HasSSE3]>; +def ISTT_Fp64m64 : FpI_<(outs), (ins i64mem:$op, RFP64:$src), OneArgFP, + [(X86fp_to_i64mem RFP64:$src, addr:$op)]>, + Requires<[HasSSE3]>; +def ISTT_Fp16m80 : FpI_<(outs), (ins i16mem:$op, RFP80:$src), OneArgFP, + [(X86fp_to_i16mem RFP80:$src, addr:$op)]>, + Requires<[HasSSE3]>; +def ISTT_Fp32m80 : FpI_<(outs), (ins i32mem:$op, RFP80:$src), OneArgFP, + [(X86fp_to_i32mem RFP80:$src, addr:$op)]>, + Requires<[HasSSE3]>; +def ISTT_Fp64m80 : FpI_<(outs), (ins i64mem:$op, RFP80:$src), OneArgFP, + [(X86fp_to_i64mem RFP80:$src, addr:$op)]>, + Requires<[HasSSE3]>; + +let mayStore = 1 in { +def ISTT_FP16m : FPI<0xDF, MRM1m, (outs), (ins i16mem:$dst), "fisttp{s}\t$dst">; +def ISTT_FP32m : FPI<0xDB, MRM1m, (outs), (ins i32mem:$dst), "fisttp{l}\t$dst">; +def ISTT_FP64m : FPI<0xDD, MRM1m, (outs), (ins i64mem:$dst), "fisttp{ll}\t$dst">; +} + +// FP Stack manipulation instructions. +def LD_Frr : FPI<0xC0, AddRegFrm, (outs), (ins RST:$op), "fld\t$op">, D9; +def ST_Frr : FPI<0xD0, AddRegFrm, (outs), (ins RST:$op), "fst\t$op">, DD; +def ST_FPrr : FPI<0xD8, AddRegFrm, (outs), (ins RST:$op), "fstp\t$op">, DD; +def XCH_F : FPI<0xC8, AddRegFrm, (outs), (ins RST:$op), "fxch\t$op">, D9; + +// Floating point constant loads. +let isReMaterializable = 1 in { +def LD_Fp032 : FpIf32<(outs RFP32:$dst), (ins), ZeroArgFP, + [(set RFP32:$dst, fpimm0)]>; +def LD_Fp132 : FpIf32<(outs RFP32:$dst), (ins), ZeroArgFP, + [(set RFP32:$dst, fpimm1)]>; +def LD_Fp064 : FpIf64<(outs RFP64:$dst), (ins), ZeroArgFP, + [(set RFP64:$dst, fpimm0)]>; +def LD_Fp164 : FpIf64<(outs RFP64:$dst), (ins), ZeroArgFP, + [(set RFP64:$dst, fpimm1)]>; +def LD_Fp080 : FpI_<(outs RFP80:$dst), (ins), ZeroArgFP, + [(set RFP80:$dst, fpimm0)]>; +def LD_Fp180 : FpI_<(outs RFP80:$dst), (ins), ZeroArgFP, + [(set RFP80:$dst, fpimm1)]>; +} + +def LD_F0 : FPI<0xEE, RawFrm, (outs), (ins), "fldz">, D9; +def LD_F1 : FPI<0xE8, RawFrm, (outs), (ins), "fld1">, D9; + + +// Floating point compares. +let Defs = [EFLAGS] in { +def UCOM_Fpr32 : FpIf32<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP, + []>; // FPSW = cmp ST(0) with ST(i) +def UCOM_Fpr64 : FpIf64<(outs), (ins RFP64:$lhs, RFP64:$rhs), CompareFP, + []>; // FPSW = cmp ST(0) with ST(i) +def UCOM_Fpr80 : FpI_ <(outs), (ins RFP80:$lhs, RFP80:$rhs), CompareFP, + []>; // FPSW = cmp ST(0) with ST(i) + +def UCOM_FpIr32: FpIf32<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP, + [(X86cmp RFP32:$lhs, RFP32:$rhs), + (implicit EFLAGS)]>; // CC = ST(0) cmp ST(i) +def UCOM_FpIr64: FpIf64<(outs), (ins RFP64:$lhs, RFP64:$rhs), CompareFP, + [(X86cmp RFP64:$lhs, RFP64:$rhs), + (implicit EFLAGS)]>; // CC = ST(0) cmp ST(i) +def UCOM_FpIr80: FpI_<(outs), (ins RFP80:$lhs, RFP80:$rhs), CompareFP, + [(X86cmp RFP80:$lhs, RFP80:$rhs), + (implicit EFLAGS)]>; // CC = ST(0) cmp ST(i) +} + +let Defs = [EFLAGS], Uses = [ST0] in { +def UCOM_Fr : FPI<0xE0, AddRegFrm, // FPSW = cmp ST(0) with ST(i) + (outs), (ins RST:$reg), + "fucom\t$reg">, DD; +def UCOM_FPr : FPI<0xE8, AddRegFrm, // FPSW = cmp ST(0) with ST(i), pop + (outs), (ins RST:$reg), + "fucomp\t$reg">, DD; +def UCOM_FPPr : FPI<0xE9, RawFrm, // cmp ST(0) with ST(1), pop, pop + (outs), (ins), + "fucompp">, DA; + +def UCOM_FIr : FPI<0xE8, AddRegFrm, // CC = cmp ST(0) with ST(i) + (outs), (ins RST:$reg), + "fucomi\t{$reg, %st(0)|%ST(0), $reg}">, DB; +def UCOM_FIPr : FPI<0xE8, AddRegFrm, // CC = cmp ST(0) with ST(i), pop + (outs), (ins RST:$reg), + "fucomip\t{$reg, %st(0)|%ST(0), $reg}">, DF; +} + +// Floating point flag ops. +let Defs = [AX] in +def FNSTSW8r : I<0xE0, RawFrm, // AX = fp flags + (outs), (ins), "fnstsw", []>, DF; + +def FNSTCW16m : I<0xD9, MRM7m, // [mem16] = X87 control world + (outs), (ins i16mem:$dst), "fnstcw\t$dst", + [(X86fp_cwd_get16 addr:$dst)]>; + +let mayLoad = 1 in +def FLDCW16m : I<0xD9, MRM5m, // X87 control world = [mem16] + (outs), (ins i16mem:$dst), "fldcw\t$dst", []>; + +//===----------------------------------------------------------------------===// +// Non-Instruction Patterns +//===----------------------------------------------------------------------===// + +// Required for RET of f32 / f64 / f80 values. +def : Pat<(X86fld addr:$src, f32), (LD_Fp32m addr:$src)>; +def : Pat<(X86fld addr:$src, f64), (LD_Fp64m addr:$src)>; +def : Pat<(X86fld addr:$src, f80), (LD_Fp80m addr:$src)>; + +// Required for CALL which return f32 / f64 / f80 values. +def : Pat<(X86fst RFP32:$src, addr:$op, f32), (ST_Fp32m addr:$op, RFP32:$src)>; +def : Pat<(X86fst RFP64:$src, addr:$op, f32), (ST_Fp64m32 addr:$op, RFP64:$src)>; +def : Pat<(X86fst RFP64:$src, addr:$op, f64), (ST_Fp64m addr:$op, RFP64:$src)>; +def : Pat<(X86fst RFP80:$src, addr:$op, f32), (ST_Fp80m32 addr:$op, RFP80:$src)>; +def : Pat<(X86fst RFP80:$src, addr:$op, f64), (ST_Fp80m64 addr:$op, RFP80:$src)>; +def : Pat<(X86fst RFP80:$src, addr:$op, f80), (ST_FpP80m addr:$op, RFP80:$src)>; + +// Floating point constant -0.0 and -1.0 +def : Pat<(f32 fpimmneg0), (CHS_Fp32 (LD_Fp032))>, Requires<[FPStackf32]>; +def : Pat<(f32 fpimmneg1), (CHS_Fp32 (LD_Fp132))>, Requires<[FPStackf32]>; +def : Pat<(f64 fpimmneg0), (CHS_Fp64 (LD_Fp064))>, Requires<[FPStackf64]>; +def : Pat<(f64 fpimmneg1), (CHS_Fp64 (LD_Fp164))>, Requires<[FPStackf64]>; +def : Pat<(f80 fpimmneg0), (CHS_Fp80 (LD_Fp080))>; +def : Pat<(f80 fpimmneg1), (CHS_Fp80 (LD_Fp180))>; + +// Used to conv. i64 to f64 since there isn't a SSE version. +def : Pat<(X86fildflag addr:$src, i64), (ILD_Fp64m64 addr:$src)>; + +// FP extensions map onto simple pseudo-value conversions if they are to/from +// the FP stack. +def : Pat<(f64 (fextend RFP32:$src)), (MOV_Fp3264 RFP32:$src)>, + Requires<[FPStackf32]>; +def : Pat<(f80 (fextend RFP32:$src)), (MOV_Fp3280 RFP32:$src)>, + Requires<[FPStackf32]>; +def : Pat<(f80 (fextend RFP64:$src)), (MOV_Fp6480 RFP64:$src)>, + Requires<[FPStackf64]>; + +// FP truncations map onto simple pseudo-value conversions if they are to/from +// the FP stack. We have validated that only value-preserving truncations make +// it through isel. +def : Pat<(f32 (fround RFP64:$src)), (MOV_Fp6432 RFP64:$src)>, + Requires<[FPStackf32]>; +def : Pat<(f32 (fround RFP80:$src)), (MOV_Fp8032 RFP80:$src)>, + Requires<[FPStackf32]>; +def : Pat<(f64 (fround RFP80:$src)), (MOV_Fp8064 RFP80:$src)>, + Requires<[FPStackf64]>; |
