X-Git-Url: http://demsky.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FTarget%2FX86%2FX86InstrFPStack.td;h=9124f90de55e0c9b9ef9c0358b17d7c73ee31909;hb=b0269cd2c8d8512ea156a6c6df798faa6c76145c;hp=5449e648cd8d8f345c132a5c5081ee505e4c3865;hpb=9b37aaf04c633c68c9ef09d3991d9acfbef9f754;p=oota-llvm.git diff --git a/lib/Target/X86/X86InstrFPStack.td b/lib/Target/X86/X86InstrFPStack.td index 5449e648cd8..9124f90de55 100644 --- a/lib/Target/X86/X86InstrFPStack.td +++ b/lib/Target/X86/X86InstrFPStack.td @@ -17,8 +17,8 @@ // FPStack specific DAG Nodes. //===----------------------------------------------------------------------===// -def SDTX86FpGet : SDTypeProfile<1, 0, [SDTCisFP<0>]>; -def SDTX86FpSet : SDTypeProfile<0, 1, [SDTCisFP<0>]>; +def SDTX86FpGet2 : SDTypeProfile<2, 0, [SDTCisVT<0, f80>, + SDTCisVT<1, f80>]>; def SDTX86Fld : SDTypeProfile<1, 2, [SDTCisFP<0>, SDTCisPtrTy<1>, SDTCisVT<2, OtherVT>]>; @@ -31,26 +31,25 @@ def SDTX86FpToIMem : SDTypeProfile<0, 2, [SDTCisFP<0>, SDTCisPtrTy<1>]>; def SDTX86CwdStore : SDTypeProfile<0, 1, [SDTCisPtrTy<0>]>; -def X86fpget : SDNode<"X86ISD::FP_GET_RESULT", SDTX86FpGet, - [SDNPHasChain, SDNPInFlag, SDNPOutFlag]>; -def X86fpset : SDNode<"X86ISD::FP_SET_RESULT", SDTX86FpSet, - [SDNPHasChain, SDNPOutFlag]>; -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 X86fld : SDNode<"X86ISD::FLD", SDTX86Fld, + [SDNPHasChain, SDNPMayLoad, SDNPMemOperand]>; +def X86fst : SDNode<"X86ISD::FST", SDTX86Fst, + [SDNPHasChain, SDNPInFlag, SDNPMayStore, + SDNPMemOperand]>; +def X86fild : SDNode<"X86ISD::FILD", SDTX86Fild, + [SDNPHasChain, SDNPMayLoad, SDNPMemOperand]>; +def X86fildflag : SDNode<"X86ISD::FILD_FLAG", SDTX86Fild, + [SDNPHasChain, SDNPOutFlag, SDNPMayLoad, + SDNPMemOperand]>; def X86fp_to_i16mem : SDNode<"X86ISD::FP_TO_INT16_IN_MEM", SDTX86FpToIMem, - [SDNPHasChain]>; + [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>; def X86fp_to_i32mem : SDNode<"X86ISD::FP_TO_INT32_IN_MEM", SDTX86FpToIMem, - [SDNPHasChain]>; + [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>; def X86fp_to_i64mem : SDNode<"X86ISD::FP_TO_INT64_IN_MEM", SDTX86FpToIMem, - [SDNPHasChain]>; + [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>; def X86fp_cwd_get16 : SDNode<"X86ISD::FNSTCW16m", SDTX86CwdStore, - [SDNPHasChain]>; + [SDNPHasChain, SDNPMayStore, SDNPSideEffect, + SDNPMemOperand]>; //===----------------------------------------------------------------------===// // FPStack pattern fragments @@ -73,49 +72,45 @@ def fpimmneg1 : PatLeaf<(fpimm), [{ }]>; // Some 'special' instructions -let usesCustomDAGSchedInserter = 1 in { // Expanded by the scheduler. +let usesCustomInserter = 1 in { // Expanded after instruction selection. def FP32_TO_INT16_IN_MEM : I<0, Pseudo, (outs), (ins i16mem:$dst, RFP32:$src), - "#FP32_TO_INT16_IN_MEM PSEUDO!", + "##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!", + "##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!", + "##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!", + "##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!", + "##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!", + "##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!", + "##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!", + "##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!", + "##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, @@ -136,27 +131,32 @@ let isTerminator = 1 in // encoding and asm printing info). // Pseudo Instructions for FP stack return values. -def FpGETRESULT32 : FpI_<(outs RFP32:$dst), (ins), SpecialFP, - [(set RFP32:$dst, X86fpget)]>; // FPR = ST(0) - -def FpGETRESULT64 : FpI_<(outs RFP64:$dst), (ins), SpecialFP, - [(set RFP64:$dst, X86fpget)]>; // FPR = ST(0) - -def FpGETRESULT80 : FpI_<(outs RFP80:$dst), (ins), SpecialFP, - [(set RFP80:$dst, X86fpget)]>; // FPR = ST(0) +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 FpSETRESULT32 : FpI_<(outs), (ins RFP32:$src), SpecialFP, - [(X86fpset RFP32:$src)]>;// ST(0) = FPR - -def FpSETRESULT64 : FpI_<(outs), (ins RFP64:$src), SpecialFP, - [(X86fpset RFP64:$src)]>;// ST(0) = FPR +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:$src), SpecialFP, []>; // ST(0) = FPR +} -def FpSETRESULT80 : FpI_<(outs), (ins RFP80:$src), SpecialFP, - [(X86fpset RFP80:$src)]>;// ST(0) = FPR +let Defs = [ST1] in { +def FpSET_ST1_32 : FpI_<(outs), (ins RFP32:$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. +// FpIf32, FpIf64 - Floating Point Pseudo 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. @@ -166,15 +166,17 @@ class FpIf64 pattern> : FpI_, Requires<[FPStackf64]>; // Register copies. Just copies, the shortening ones do not truncate. -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, []>; +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 { @@ -192,48 +194,67 @@ def _Fp80 : FpI_<(outs RFP80:$dst), (ins RFP80:$src1, RFP80:$src2), TwoArgFP, // These instructions cannot address 80-bit memory. multiclass FPBinary { // ST(0) = ST(0) + [mem] -def _Fp32m : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, f32mem:$src2), OneArgFPRW, +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:$src2), OneArgFPRW, +def _Fp64m : FpIf64<(outs RFP64:$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, +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:$src2), OneArgFPRW, +def _Fp80m32: FpI_<(outs RFP80:$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, +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"))>; + !strconcat("f", asmstring, "{s}\t$src")> { + let mayLoad = 1; +} def _F64m : FPI<0xDC, fp, (outs), (ins f64mem:$src), - !strconcat("f", !strconcat(asmstring, "{l}\t$src"))>; + !strconcat("f", asmstring, "{l}\t$src")> { + let mayLoad = 1; +} // ST(0) = ST(0) + [memint] -def _FpI16m32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, i16mem:$src2), OneArgFPRW, +def _FpI16m32 : FpIf32<(outs RFP32:$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, +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:$src2), OneArgFPRW, +def _FpI16m64 : FpIf64<(outs RFP64:$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, +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:$src2), OneArgFPRW, +def _FpI16m80 : FpI_<(outs RFP80:$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, +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"))>; + !strconcat("fi", asmstring, "{s}\t$src")> { + let mayLoad = 1; +} def _FI32m : FPI<0xDA, fp, (outs), (ins i32mem:$src), - !strconcat("fi", !strconcat(asmstring, "{l}\t$src"))>; + !strconcat("fi", asmstring, "{l}\t$src")> { + let mayLoad = 1; +} } defm ADD : FPBinary_rr; @@ -276,6 +297,9 @@ 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$op">; +def COM_FST0r : FPST0rInst <0xD0, "fcom\t$op">; +def COMP_FST0r : FPST0rInst <0xD8, "fcomp\t$op">; + // Unary operations. multiclass FPUnary opcode, string asmstring> { def _Fp32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src), OneArgFPRW, @@ -293,30 +317,60 @@ defm SQRT: FPUnary; defm SIN : FPUnary; defm COS : FPUnary; -def TST_Fp32 : FpIf32<(outs), (ins RFP32:$src), OneArgFP, - []>; -def TST_Fp64 : FpIf64<(outs), (ins RFP64:$src), OneArgFP, - []>; -def TST_Fp80 : FpI_<(outs), (ins RFP80:$src), OneArgFP, - []>; +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:$src), OneArgFP, []>; +} def TST_F : FPI<0xE4, RawFrm, (outs), (ins), "ftst">, D9; +// Versions of FP instructions that take a single memory operand. Added for the +// disassembler; remove as they are included with patterns elsewhere. +def FCOM32m : FPI<0xD8, MRM2m, (outs), (ins f32mem:$src), "fcom{s}\t$src">; +def FCOMP32m : FPI<0xD8, MRM3m, (outs), (ins f32mem:$src), "fcomp{s}\t$src">; + +def FLDENVm : FPI<0xD9, MRM4m, (outs), (ins f32mem:$src), "fldenv\t$src">; +def FSTENVm : FPI<0xD9, MRM6m, (outs f32mem:$dst), (ins), "fnstenv\t$dst">; + +def FICOM32m : FPI<0xDA, MRM2m, (outs), (ins i32mem:$src), "ficom{l}\t$src">; +def FICOMP32m: FPI<0xDA, MRM3m, (outs), (ins i32mem:$src), "ficomp{l}\t$src">; + +def FCOM64m : FPI<0xDC, MRM2m, (outs), (ins f64mem:$src), "fcom{l}\t$src">; +def FCOMP64m : FPI<0xDC, MRM3m, (outs), (ins f64mem:$src), "fcomp{l}\t$src">; + +def FRSTORm : FPI<0xDD, MRM4m, (outs f32mem:$dst), (ins), "frstor\t$dst">; +def FSAVEm : FPI<0xDD, MRM6m, (outs f32mem:$dst), (ins), "fnsave\t$dst">; +def FNSTSWm : FPI<0xDD, MRM7m, (outs f32mem:$dst), (ins), "fnstsw\t$dst">; + +def FICOM16m : FPI<0xDE, MRM2m, (outs), (ins i16mem:$src), "ficom{s}\t$src">; +def FICOMP16m: FPI<0xDE, MRM3m, (outs), (ins i16mem:$src), "ficomp{s}\t$src">; + +def FBLDm : FPI<0xDF, MRM4m, (outs), (ins f32mem:$src), "fbld\t$src">; +def FBSTPm : FPI<0xDF, MRM6m, (outs f32mem:$dst), (ins), "fbstp\t$dst">; + // Floating point cmovs. +class FpIf32CMov pattern> : + FpI_, Requires<[FPStackf32, HasCMov]>; +class FpIf64CMov pattern> : + FpI_, Requires<[FPStackf64, HasCMov]>; + multiclass FPCMov { - def _Fp32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, RFP32:$src2), + def _Fp32 : FpIf32CMov<(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:$src2), + def _Fp64 : FpIf64CMov<(outs RFP64:$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:$src2, - cc, EFLAGS))]>; + cc, EFLAGS))]>, + Requires<[HasCMov]>; } -let Uses = [EFLAGS], isTwoAddress = 1 in { + +let Uses = [EFLAGS], Constraints = "$src1 = $dst" in { defm CMOVB : FPCMov; defm CMOVBE : FPCMov; defm CMOVE : FPCMov; @@ -325,8 +379,9 @@ defm CMOVNB : FPCMov; defm CMOVNBE: FPCMov; defm CMOVNE : FPCMov; defm CMOVNP : FPCMov; -} +} // Uses = [EFLAGS], Constraints = "$src1 = $dst" +let Predicates = [HasCMov] in { // These are not factored because there's no clean way to pass DA/DB. def CMOVB_F : FPI<0xC0, AddRegFrm, (outs RST:$op), (ins), "fcmovb\t{$op, %st(0)|%ST(0), $op}">, DA; @@ -344,12 +399,13 @@ def CMOVNE_F : FPI<0xC8, AddRegFrm, (outs RST:$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), $op}">, DB; +} // Predicates = [HasCMov] // Floating point loads & stores. -let isSimpleLoad = 1 in { +let canFoldAsLoad = 1 in { def LD_Fp32m : FpIf32<(outs RFP32:$dst), (ins f32mem:$src), ZeroArgFP, [(set RFP32:$dst, (loadf32 addr:$src))]>; -let isReMaterializable = 1, mayHaveSideEffects = 1 in +let isReMaterializable = 1 in def LD_Fp64m : FpIf64<(outs RFP64:$dst), (ins f64mem:$src), ZeroArgFP, [(set RFP64:$dst, (loadf64 addr:$src))]>; def LD_Fp80m : FpI_<(outs RFP80:$dst), (ins f80mem:$src), ZeroArgFP, @@ -392,13 +448,16 @@ def ST_Fp80m64 : FpI_<(outs), (ins f64mem:$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:$src), OneArgFP, []>; @@ -408,13 +467,17 @@ def IST_Fp64m64 : FpIf64<(outs), (ins i64mem:$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$dst">; @@ -425,6 +488,7 @@ def IST_F32m : FPI<0xDB, MRM2m, (outs), (ins i32mem:$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:$src), OneArgFP, @@ -455,9 +519,12 @@ def ISTT_Fp64m80 : FpI_<(outs), (ins i64mem:$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$dst">; +def ISTT_FP64m : FPI<0xDD, MRM1m, (outs), (ins i64mem:$dst), + "fisttp{ll}\t$dst">; +} // FP Stack manipulation instructions. def LD_Frr : FPI<0xC0, AddRegFrm, (outs), (ins RST:$op), "fld\t$op">, D9; @@ -466,7 +533,7 @@ def ST_FPrr : FPI<0xD8, AddRegFrm, (outs), (ins RST:$op), "fstp\t$op">, DD; def XCH_F : FPI<0xC8, AddRegFrm, (outs), (ins RST:$op), "fxch\t$op">, D9; // Floating point constant loads. -let isReMaterializable = 1, neverHasSideEffects = 1 in { +let isReMaterializable = 1 in { def LD_Fp032 : FpIf32<(outs RFP32:$dst), (ins), ZeroArgFP, [(set RFP32:$dst, fpimm0)]>; def LD_Fp132 : FpIf32<(outs RFP32:$dst), (ins), ZeroArgFP, @@ -488,20 +555,19 @@ 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_FpIr32: FpIf32<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP, - [(X86cmp RFP32:$lhs, RFP32:$rhs), - (implicit EFLAGS)]>; // CC = ST(0) cmp ST(i) + []>; // 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) + []>; // 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) + +// CC = ST(0) cmp ST(i) +def UCOM_FpIr32: FpIf32<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP, + [(set EFLAGS, (X86cmp RFP32:$lhs, RFP32:$rhs))]>; 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_Fpr80 : FpI_<(outs), (ins RFP80:$lhs, RFP80:$rhs), CompareFP, - []>; // FPSW = cmp ST(0) with ST(i) + [(set EFLAGS, (X86cmp RFP64:$lhs, RFP64:$rhs))]>; def UCOM_FpIr80: FpI_<(outs), (ins RFP80:$lhs, RFP80:$rhs), CompareFP, - [(X86cmp RFP80:$lhs, RFP80:$rhs), - (implicit EFLAGS)]>; // CC = ST(0) cmp ST(i) + [(set EFLAGS, (X86cmp RFP80:$lhs, RFP80:$rhs))]>; } let Defs = [EFLAGS], Uses = [ST0] in { @@ -517,23 +583,69 @@ def UCOM_FPPr : FPI<0xE9, RawFrm, // cmp ST(0) with ST(1), pop, pop def UCOM_FIr : FPI<0xE8, AddRegFrm, // CC = cmp ST(0) with ST(i) (outs), (ins RST:$reg), - "fucomi\t{$reg, %st(0)|%ST(0), $reg}">, DB; + "fucomi\t$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), $reg}">, DF; + "fucompi\t$reg">, DF; } +def COM_FIr : FPI<0xF0, AddRegFrm, (outs), (ins RST:$reg), + "fcomi\t$reg">, DB; +def COM_FIPr : FPI<0xF0, AddRegFrm, (outs), (ins RST:$reg), + "fcompi\t$reg">, DF; + // Floating point flag ops. let Defs = [AX] in def FNSTSW8r : I<0xE0, RawFrm, // AX = fp flags - (outs), (ins), "fnstsw", []>, DF; + (outs), (ins), "fnstsw %ax", []>, 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$dst", []>; +// FPU control instructions +def FNINIT : I<0xE3, RawFrm, (outs), (ins), "fninit", []>, DB; +def FFREE : FPI<0xC0, AddRegFrm, (outs), (ins RST:$reg), + "ffree\t$reg">, DD; + +// Clear exceptions + +def FNCLEX : I<0xE2, RawFrm, (outs), (ins), "fnclex", []>, DB; + +// Operandless floating-point instructions for the disassembler. +def WAIT : I<0x9B, RawFrm, (outs), (ins), "wait", []>; + +def FNOP : I<0xD0, RawFrm, (outs), (ins), "fnop", []>, D9; +def FXAM : I<0xE5, RawFrm, (outs), (ins), "fxam", []>, D9; +def FLDL2T : I<0xE9, RawFrm, (outs), (ins), "fldl2t", []>, D9; +def FLDL2E : I<0xEA, RawFrm, (outs), (ins), "fldl2e", []>, D9; +def FLDPI : I<0xEB, RawFrm, (outs), (ins), "fldpi", []>, D9; +def FLDLG2 : I<0xEC, RawFrm, (outs), (ins), "fldlg2", []>, D9; +def FLDLN2 : I<0xED, RawFrm, (outs), (ins), "fldln2", []>, D9; +def F2XM1 : I<0xF0, RawFrm, (outs), (ins), "f2xm1", []>, D9; +def FYL2X : I<0xF1, RawFrm, (outs), (ins), "fyl2x", []>, D9; +def FPTAN : I<0xF2, RawFrm, (outs), (ins), "fptan", []>, D9; +def FPATAN : I<0xF3, RawFrm, (outs), (ins), "fpatan", []>, D9; +def FXTRACT : I<0xF4, RawFrm, (outs), (ins), "fxtract", []>, D9; +def FPREM1 : I<0xF5, RawFrm, (outs), (ins), "fprem1", []>, D9; +def FDECSTP : I<0xF6, RawFrm, (outs), (ins), "fdecstp", []>, D9; +def FINCSTP : I<0xF7, RawFrm, (outs), (ins), "fincstp", []>, D9; +def FPREM : I<0xF8, RawFrm, (outs), (ins), "fprem", []>, D9; +def FYL2XP1 : I<0xF9, RawFrm, (outs), (ins), "fyl2xp1", []>, D9; +def FSINCOS : I<0xFB, RawFrm, (outs), (ins), "fsincos", []>, D9; +def FRNDINT : I<0xFC, RawFrm, (outs), (ins), "frndint", []>, D9; +def FSCALE : I<0xFD, RawFrm, (outs), (ins), "fscale", []>, D9; +def FCOMPP : I<0xD9, RawFrm, (outs), (ins), "fcompp", []>, DE; + +def FXSAVE : I<0xAE, MRM0m, (outs opaque512mem:$dst), (ins), + "fxsave\t$dst", []>, TB; +def FXRSTOR : I<0xAE, MRM1m, (outs), (ins opaque512mem:$src), + "fxrstor\t$src", []>, TB; + //===----------------------------------------------------------------------===// // Non-Instruction Patterns //===----------------------------------------------------------------------===// @@ -545,11 +657,15 @@ def : Pat<(X86fld addr:$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:$op, f32), (ST_Fp64m32 addr:$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:$src)>; +def : Pat<(X86fst RFP80:$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:$src)>; // Floating point constant -0.0 and -1.0 def : Pat<(f32 fpimmneg0), (CHS_Fp32 (LD_Fp032))>, Requires<[FPStackf32]>; @@ -562,6 +678,21 @@ def : Pat<(f80 fpimmneg1), (CHS_Fp80 (LD_Fp180))>; // Used to conv. i64 to f64 since there isn't a SSE version. def : Pat<(X86fildflag addr:$src, i64), (ILD_Fp64m64 addr:$src)>; -def : Pat<(f64 (fextend RFP32:$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:$src)>, Requires<[FPStackf64]>; +// FP extensions map onto simple pseudo-value conversions if they are to/from +// the FP stack. +def : Pat<(f64 (fextend RFP32:$src)), (COPY_TO_REGCLASS RFP32:$src, RFP64)>, + Requires<[FPStackf32]>; +def : Pat<(f80 (fextend RFP32:$src)), (COPY_TO_REGCLASS RFP32:$src, RFP80)>, + Requires<[FPStackf32]>; +def : Pat<(f80 (fextend RFP64:$src)), (COPY_TO_REGCLASS RFP64:$src, RFP80)>, + Requires<[FPStackf64]>; + +// FP truncations map onto simple pseudo-value conversions if they are to/from +// the FP stack. We have validated that only value-preserving truncations make +// it through isel. +def : Pat<(f32 (fround RFP64:$src)), (COPY_TO_REGCLASS RFP64:$src, RFP32)>, + Requires<[FPStackf32]>; +def : Pat<(f32 (fround RFP80:$src)), (COPY_TO_REGCLASS RFP80:$src, RFP32)>, + Requires<[FPStackf32]>; +def : Pat<(f64 (fround RFP80:$src)), (COPY_TO_REGCLASS RFP80:$src, RFP64)>, + Requires<[FPStackf64]>;