diff --git a/llvm/lib/Target/LoongArch/LoongArch.td b/llvm/lib/Target/LoongArch/LoongArch.td --- a/llvm/lib/Target/LoongArch/LoongArch.td +++ b/llvm/lib/Target/LoongArch/LoongArch.td @@ -132,9 +132,7 @@ //===----------------------------------------------------------------------===// def LoongArchInstrInfo : InstrInfo { - // guess mayLoad, mayStore, and hasSideEffects - // This option is a temporary migration help. It will go away. - let guessInstructionProperties = 1; + let guessInstructionProperties = 0; } def LoongArchAsmParser : AsmParser { diff --git a/llvm/lib/Target/LoongArch/LoongArchFloat32InstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchFloat32InstrInfo.td --- a/llvm/lib/Target/LoongArch/LoongArchFloat32InstrInfo.td +++ b/llvm/lib/Target/LoongArch/LoongArchFloat32InstrInfo.td @@ -98,8 +98,10 @@ def FMOV_S : FP_MOV<0b0000000100010100100101, "fmov.s", FPR32, FPR32>; def MOVGR2FR_W : FP_MOV<0b0000000100010100101001, "movgr2fr.w", FPR32, GPR>; def MOVFR2GR_S : FP_MOV<0b0000000100010100101101, "movfr2gr.s", GPR, FPR32>; +let hasSideEffects = 1 in { def MOVGR2FCSR : FP_MOV<0b0000000100010100110000, "movgr2fcsr", FCSR, GPR>; def MOVFCSR2GR : FP_MOV<0b0000000100010100110010, "movfcsr2gr", GPR, FCSR>; +} // hasSideEffects = 1 def MOVFR2CF_S : FP_MOV<0b0000000100010100110100, "movfr2cf", CFR, FPR32>; def MOVCF2FR_S : FP_MOV<0b0000000100010100110101, "movcf2fr", FPR32, CFR>; def MOVGR2CF : FP_MOV<0b0000000100010100110110, "movgr2cf", CFR, GPR>; diff --git a/llvm/lib/Target/LoongArch/LoongArchFloat64InstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchFloat64InstrInfo.td --- a/llvm/lib/Target/LoongArch/LoongArchFloat64InstrInfo.td +++ b/llvm/lib/Target/LoongArch/LoongArchFloat64InstrInfo.td @@ -105,11 +105,11 @@ def MOVFR2GR_S_64 : FP_MOV<0b0000000100010100101101, "movfr2gr.s", GPR, FPR64>; def FSEL_D : FP_SEL<0b00001101000000, "fsel", FPR64>; } // isCodeGenOnly = 1 -let Constraints = "$dst = $out" in { +let hasSideEffects = 0, mayLoad = 0, mayStore = 0, Constraints = "$dst = $out" in { def MOVGR2FRH_W : FPFmtMOV<0b0000000100010100101011, (outs FPR64:$out), (ins FPR64:$dst, GPR:$src), "movgr2frh.w", "$dst, $src">; -} // Constraints = "$dst = $out" +} // hasSideEffects = 0, mayLoad = 0, mayStore = 0, Constraints = "$dst = $out" // Common Memory Access Instructions def FLD_D : FP_LOAD_2RI12<0b0010101110, "fld.d", FPR64>; diff --git a/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td b/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td --- a/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td +++ b/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td @@ -157,6 +157,7 @@ // Instruction class templates //===----------------------------------------------------------------------===// +let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in { class FP_ALU_2R op, string opstr, RegisterClass rc> : FPFmt2R; @@ -166,6 +167,7 @@ class FP_ALU_4R op, string opstr, RegisterClass rc> : FPFmt4R; +} // hasSideEffects = 0, mayLoad = 0, mayStore = 0 class FPCMPOpc value> { bits<12> val = value; @@ -175,6 +177,7 @@ bits<5> val = value; } +let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in { class FP_CMP : FPFmtFCMP; @@ -195,24 +198,25 @@ let isBranch = 1; let isTerminator = 1; } +} // hasSideEffects = 0, mayLoad = 0, mayStore = 0 -let mayLoad = 1 in { +let hasSideEffects = 0, mayLoad = 1, mayStore = 0 in { class FP_LOAD_3R op, string opstr, RegisterClass rc> : FPFmtMEM; class FP_LOAD_2RI12 op, string opstr, RegisterClass rc> : FPFmt2RI12; -} // mayLoad = 1 +} // hasSideEffects = 0, mayLoad = 1, mayStore = 0 -let mayStore = 1 in { +let hasSideEffects = 0, mayLoad = 0, mayStore = 1 in { class FP_STORE_3R op, string opstr, RegisterClass rc> : FPFmtMEM; class FP_STORE_2RI12 op, string opstr, RegisterClass rc> : FPFmt2RI12; -} // mayStore = 1 +} // hasSideEffects = 0, mayLoad = 0, mayStore = 1 def FPCMP_OPC_S : FPCMPOpc<0b000011000001>; def FPCMP_OPC_D : FPCMPOpc<0b000011000010>; diff --git a/llvm/lib/Target/LoongArch/LoongArchInstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchInstrInfo.td --- a/llvm/lib/Target/LoongArch/LoongArchInstrInfo.td +++ b/llvm/lib/Target/LoongArch/LoongArchInstrInfo.td @@ -473,6 +473,7 @@ // Instruction Class Templates //===----------------------------------------------------------------------===// +let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in { class ALU_3R op, string opstr> : Fmt3R; class ALU_2R op, string opstr> @@ -498,13 +499,17 @@ "$rd, $rj, $imm16">; class ALU_1RI20 op, string opstr, Operand ImmOpnd> : Fmt1RI20; +} // hasSideEffects = 0, mayLoad = 0, mayStore = 0 +let hasSideEffects = 1, mayLoad = 0, mayStore = 0 in class MISC_I15 op, string opstr> : FmtI15; +let hasSideEffects = 1, mayLoad = 0, mayStore = 0 in class RDTIME_2R op, string opstr> : Fmt2R; +let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in { class BrCC_2RI16 op, string opstr> : Fmt2RI16 { @@ -522,8 +527,9 @@ let isBranch = 1; let isTerminator = 1; } +} // hasSideEffects = 0, mayLoad = 0, mayStore = 0 -let mayLoad = 1 in { +let hasSideEffects = 0, mayLoad = 1, mayStore = 0 in { class LOAD_3R op, string opstr> : Fmt3R; class LOAD_2RI12 op, string opstr> @@ -532,9 +538,9 @@ class LOAD_2RI14 op, string opstr> : Fmt2RI14; -} // mayLoad = 1 +} // hasSideEffects = 0, mayLoad = 1, mayStore = 0 -let mayStore = 1 in { +let hasSideEffects = 0, mayLoad = 0, mayStore = 1 in { class STORE_3R op, string opstr> : Fmt3R; @@ -544,26 +550,28 @@ class STORE_2RI14 op, string opstr> : Fmt2RI14; -} // mayStore = 1 +} // hasSideEffects = 0, mayLoad = 0, mayStore = 1 -let mayLoad = 1, mayStore = 1, Constraints = "@earlyclobber $rd" in +let hasSideEffects = 0, mayLoad = 1, mayStore = 1, Constraints = "@earlyclobber $rd" in class AM_3R op, string opstr> : Fmt3R; -let mayLoad = 1 in +let hasSideEffects = 0, mayLoad = 1, mayStore = 0 in class LLBase op, string opstr> : Fmt2RI14; -let mayStore = 1, Constraints = "$rd = $dst" in +let hasSideEffects = 0, mayLoad = 0, mayStore = 1, Constraints = "$rd = $dst" in class SCBase op, string opstr> : Fmt2RI14; +let hasSideEffects = 1 in class IOCSRRD op, string opstr> : Fmt2R; +let hasSideEffects = 1 in class IOCSRWR op, string opstr> : Fmt2R; @@ -647,8 +655,9 @@ def BNEZ : BrCCZ_1RI21<0b010001, "bnez">; def B : Br_I26<0b010100, "b">; -let isCall = 1, Defs=[R1] in +let hasSideEffects = 0, mayLoad = 0, mayStore = 0, isCall = 1, Defs=[R1] in def BL : FmtI26<0b010101, (outs), (ins simm26_symbol:$imm26), "bl", "$imm26">; +let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in def JIRL : Fmt2RI16<0b010011, (outs GPR:$rd), (ins GPR:$rj, simm16_lsl2:$imm16), "jirl", "$rd, $rj, $imm16">; @@ -662,6 +671,7 @@ def ST_B : STORE_2RI12<0b0010100100, "st.b">; def ST_H : STORE_2RI12<0b0010100101, "st.h">; def ST_W : STORE_2RI12<0b0010100110, "st.w">; +let hasSideEffects = 0, mayLoad = 1, mayStore = 1 in def PRELD : FmtPRELD<(outs), (ins uimm5:$imm5, GPR:$rj, simm12:$imm12), "preld", "$imm5, $rj, $imm12">; @@ -696,6 +706,7 @@ def ALSL_WU : ALU_3RI2<0b000000000000011, "alsl.wu", uimm2_plus1>; def ALSL_D : ALU_3RI2<0b000000000010110, "alsl.d", uimm2_plus1>; let Constraints = "$rd = $dst" in { +let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in def LU32I_D : Fmt1RI20<0b0001011, (outs GPR:$dst), (ins GPR:$rd, simm20_lu32id:$imm20), "lu32i.d", "$rd, $imm20">; @@ -765,6 +776,7 @@ def LDPTR_D : LOAD_2RI14<0b00100110, "ldptr.d">; def STPTR_W : STORE_2RI14<0b00100101, "stptr.w">; def STPTR_D : STORE_2RI14<0b00100111, "stptr.d">; +let hasSideEffects = 0, mayLoad = 1, mayStore = 1 in def PRELDX : FmtPRELDX<(outs), (ins uimm5:$imm5, GPR:$rj, GPR:$rk), "preldx", "$imm5, $rj, $rk">; @@ -1262,7 +1274,7 @@ [(loongarch_call GPR:$rj)]>, PseudoInstExpansion<(JIRL R1, GPR:$rj, 0)>; -let isCall = 1, Defs = [R1] in +let isCall = 1, hasSideEffects = 0, mayStore = 0, mayLoad = 0, Defs = [R1] in def PseudoJIRL_CALL : Pseudo<(outs), (ins GPR:$rj, simm16_lsl2:$imm16)>, PseudoInstExpansion<(JIRL R1, GPR:$rj, simm16_lsl2:$imm16)>; @@ -1284,11 +1296,13 @@ [(loongarch_tail GPRT:$rj)]>, PseudoInstExpansion<(JIRL R0, GPR:$rj, 0)>; -let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1, Uses = [R3] in +let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1, + hasSideEffects = 0, mayStore = 0, mayLoad = 0, Uses = [R3] in def PseudoB_TAIL : Pseudo<(outs), (ins simm26_b:$imm26)>, PseudoInstExpansion<(B simm26_b:$imm26)>; -let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1, Uses = [R3] in +let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1, + hasSideEffects = 0, mayStore = 0, mayLoad = 0, Uses = [R3] in def PseudoJIRL_TAIL : Pseudo<(outs), (ins GPR:$rj, simm16_lsl2:$imm16)>, PseudoInstExpansion<(JIRL R0, GPR:$rj, simm16_lsl2:$imm16)>; @@ -1875,15 +1889,16 @@ //===----------------------------------------------------------------------===// // CSR Access Instructions +let hasSideEffects = 1 in def CSRRD : FmtCSR<0b0000010000000, (outs GPR:$rd), (ins uimm14:$csr_num), "csrrd", "$rd, $csr_num">; -let Constraints = "$rd = $dst" in { +let hasSideEffects = 1, Constraints = "$rd = $dst" in { def CSRWR : FmtCSR<0b0000010000001, (outs GPR:$dst), (ins GPR:$rd, uimm14:$csr_num), "csrwr", "$rd, $csr_num">; def CSRXCHG : FmtCSRXCHG<0b00000100, (outs GPR:$dst), (ins GPR:$rd, GPR:$rj, uimm14:$csr_num), "csrxchg", "$rd, $rj, $csr_num">; -} // Constraints = "$rd = $dst" +} // hasSideEffects = 1, Constraints = "$rd = $dst" // IOCSR Access Instructions def IOCSRRD_B : IOCSRRD<0b0000011001001000000000, "iocsrrd.b">; @@ -1898,6 +1913,7 @@ } // Predicates = [IsLA64] // TLB Maintenance Instructions +let hasSideEffects = 1, mayLoad = 0, mayStore = 0 in { def TLBSRCH : FmtI32<0b00000110010010000010100000000000, "tlbsrch">; def TLBRD : FmtI32<0b00000110010010000010110000000000, "tlbrd">; def TLBWR : FmtI32<0b00000110010010000011000000000000, "tlbwr">; @@ -1906,6 +1922,7 @@ def TLBFLUSH : FmtI32<0b00000110010010000010010000000000, "tlbflush">; def INVTLB : FmtINVTLB<(outs), (ins GPR:$rk, GPR:$rj, uimm5:$op), "invtlb", "$op, $rj, $rk">; +} // hasSideEffects = 1, mayLoad = 0, mayStore = 0 // Software Page Walking Instructions def LDDIR : Fmt2RI8<0b00000110010000, (outs GPR:$rd), @@ -1914,6 +1931,7 @@ // Other Miscellaneous Instructions +let hasSideEffects = 1, mayLoad = 0, mayStore = 0 in def ERTN : FmtI32<0b00000110010010000011100000000000, "ertn">; def DBCL : MISC_I15<0b00000000001010101, "dbcl">; def IDLE : MISC_I15<0b00000110010010001, "idle">; diff --git a/llvm/test/CodeGen/LoongArch/ir-instruction/atomicrmw-fp.ll b/llvm/test/CodeGen/LoongArch/ir-instruction/atomicrmw-fp.ll --- a/llvm/test/CodeGen/LoongArch/ir-instruction/atomicrmw-fp.ll +++ b/llvm/test/CodeGen/LoongArch/ir-instruction/atomicrmw-fp.ll @@ -10,11 +10,11 @@ ; LA64F-NEXT: fld.s $fa0, $a0, 0 ; LA64F-NEXT: addi.w $a1, $zero, 1 ; LA64F-NEXT: movgr2fr.w $fa1, $a1 +; LA64F-NEXT: ffint.s.w $fa1, $fa1 ; LA64F-NEXT: .LBB0_1: # %atomicrmw.start ; LA64F-NEXT: # =>This Loop Header: Depth=1 ; LA64F-NEXT: # Child Loop BB0_3 Depth 2 -; LA64F-NEXT: ffint.s.w $fa2, $fa1 -; LA64F-NEXT: fadd.s $fa2, $fa0, $fa2 +; LA64F-NEXT: fadd.s $fa2, $fa0, $fa1 ; LA64F-NEXT: movfr2gr.s $a1, $fa2 ; LA64F-NEXT: movfr2gr.s $a2, $fa0 ; LA64F-NEXT: .LBB0_3: # %atomicrmw.start @@ -45,11 +45,11 @@ ; LA64D-NEXT: fld.s $fa0, $a0, 0 ; LA64D-NEXT: addi.w $a1, $zero, 1 ; LA64D-NEXT: movgr2fr.w $fa1, $a1 +; LA64D-NEXT: ffint.s.w $fa1, $fa1 ; LA64D-NEXT: .LBB0_1: # %atomicrmw.start ; LA64D-NEXT: # =>This Loop Header: Depth=1 ; LA64D-NEXT: # Child Loop BB0_3 Depth 2 -; LA64D-NEXT: ffint.s.w $fa2, $fa1 -; LA64D-NEXT: fadd.s $fa2, $fa0, $fa2 +; LA64D-NEXT: fadd.s $fa2, $fa0, $fa1 ; LA64D-NEXT: movfr2gr.s $a1, $fa2 ; LA64D-NEXT: movfr2gr.s $a2, $fa0 ; LA64D-NEXT: .LBB0_3: # %atomicrmw.start @@ -158,12 +158,12 @@ ; LA64F-NEXT: fld.s $fa0, $a0, 0 ; LA64F-NEXT: addi.w $a1, $zero, 1 ; LA64F-NEXT: movgr2fr.w $fa1, $a1 +; LA64F-NEXT: ffint.s.w $fa1, $fa1 ; LA64F-NEXT: .LBB2_1: # %atomicrmw.start ; LA64F-NEXT: # =>This Loop Header: Depth=1 ; LA64F-NEXT: # Child Loop BB2_3 Depth 2 -; LA64F-NEXT: ffint.s.w $fa2, $fa1 -; LA64F-NEXT: fmax.s $fa3, $fa0, $fa0 -; LA64F-NEXT: fmin.s $fa2, $fa3, $fa2 +; LA64F-NEXT: fmax.s $fa2, $fa0, $fa0 +; LA64F-NEXT: fmin.s $fa2, $fa2, $fa1 ; LA64F-NEXT: movfr2gr.s $a1, $fa2 ; LA64F-NEXT: movfr2gr.s $a2, $fa0 ; LA64F-NEXT: .LBB2_3: # %atomicrmw.start @@ -194,12 +194,12 @@ ; LA64D-NEXT: fld.s $fa0, $a0, 0 ; LA64D-NEXT: addi.w $a1, $zero, 1 ; LA64D-NEXT: movgr2fr.w $fa1, $a1 +; LA64D-NEXT: ffint.s.w $fa1, $fa1 ; LA64D-NEXT: .LBB2_1: # %atomicrmw.start ; LA64D-NEXT: # =>This Loop Header: Depth=1 ; LA64D-NEXT: # Child Loop BB2_3 Depth 2 -; LA64D-NEXT: ffint.s.w $fa2, $fa1 -; LA64D-NEXT: fmax.s $fa3, $fa0, $fa0 -; LA64D-NEXT: fmin.s $fa2, $fa3, $fa2 +; LA64D-NEXT: fmax.s $fa2, $fa0, $fa0 +; LA64D-NEXT: fmin.s $fa2, $fa2, $fa1 ; LA64D-NEXT: movfr2gr.s $a1, $fa2 ; LA64D-NEXT: movfr2gr.s $a2, $fa0 ; LA64D-NEXT: .LBB2_3: # %atomicrmw.start @@ -234,12 +234,12 @@ ; LA64F-NEXT: fld.s $fa0, $a0, 0 ; LA64F-NEXT: addi.w $a1, $zero, 1 ; LA64F-NEXT: movgr2fr.w $fa1, $a1 +; LA64F-NEXT: ffint.s.w $fa1, $fa1 ; LA64F-NEXT: .LBB3_1: # %atomicrmw.start ; LA64F-NEXT: # =>This Loop Header: Depth=1 ; LA64F-NEXT: # Child Loop BB3_3 Depth 2 -; LA64F-NEXT: ffint.s.w $fa2, $fa1 -; LA64F-NEXT: fmax.s $fa3, $fa0, $fa0 -; LA64F-NEXT: fmax.s $fa2, $fa3, $fa2 +; LA64F-NEXT: fmax.s $fa2, $fa0, $fa0 +; LA64F-NEXT: fmax.s $fa2, $fa2, $fa1 ; LA64F-NEXT: movfr2gr.s $a1, $fa2 ; LA64F-NEXT: movfr2gr.s $a2, $fa0 ; LA64F-NEXT: .LBB3_3: # %atomicrmw.start @@ -270,12 +270,12 @@ ; LA64D-NEXT: fld.s $fa0, $a0, 0 ; LA64D-NEXT: addi.w $a1, $zero, 1 ; LA64D-NEXT: movgr2fr.w $fa1, $a1 +; LA64D-NEXT: ffint.s.w $fa1, $fa1 ; LA64D-NEXT: .LBB3_1: # %atomicrmw.start ; LA64D-NEXT: # =>This Loop Header: Depth=1 ; LA64D-NEXT: # Child Loop BB3_3 Depth 2 -; LA64D-NEXT: ffint.s.w $fa2, $fa1 -; LA64D-NEXT: fmax.s $fa3, $fa0, $fa0 -; LA64D-NEXT: fmax.s $fa2, $fa3, $fa2 +; LA64D-NEXT: fmax.s $fa2, $fa0, $fa0 +; LA64D-NEXT: fmax.s $fa2, $fa2, $fa1 ; LA64D-NEXT: movfr2gr.s $a1, $fa2 ; LA64D-NEXT: movfr2gr.s $a2, $fa0 ; LA64D-NEXT: .LBB3_3: # %atomicrmw.start @@ -307,43 +307,46 @@ define double @double_fadd_acquire(ptr %p) nounwind { ; LA64F-LABEL: double_fadd_acquire: ; LA64F: # %bb.0: -; LA64F-NEXT: addi.d $sp, $sp, -64 -; LA64F-NEXT: st.d $ra, $sp, 56 # 8-byte Folded Spill -; LA64F-NEXT: st.d $fp, $sp, 48 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s0, $sp, 40 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s1, $sp, 32 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s2, $sp, 24 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s3, $sp, 16 # 8-byte Folded Spill +; LA64F-NEXT: addi.d $sp, $sp, -80 +; LA64F-NEXT: st.d $ra, $sp, 72 # 8-byte Folded Spill +; LA64F-NEXT: st.d $fp, $sp, 64 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s0, $sp, 56 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s1, $sp, 48 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s2, $sp, 40 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s3, $sp, 32 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s4, $sp, 24 # 8-byte Folded Spill ; LA64F-NEXT: move $fp, $a0 ; LA64F-NEXT: ld.d $a0, $a0, 0 -; LA64F-NEXT: ori $s0, $zero, 8 -; LA64F-NEXT: addi.d $s1, $sp, 8 -; LA64F-NEXT: addi.d $s2, $sp, 0 -; LA64F-NEXT: ori $s3, $zero, 2 +; LA64F-NEXT: lu52i.d $s0, $zero, 1023 +; LA64F-NEXT: ori $s1, $zero, 8 +; LA64F-NEXT: addi.d $s2, $sp, 16 +; LA64F-NEXT: addi.d $s3, $sp, 8 +; LA64F-NEXT: ori $s4, $zero, 2 ; LA64F-NEXT: .LBB4_1: # %atomicrmw.start ; LA64F-NEXT: # =>This Inner Loop Header: Depth=1 -; LA64F-NEXT: st.d $a0, $sp, 8 -; LA64F-NEXT: lu52i.d $a1, $zero, 1023 +; LA64F-NEXT: st.d $a0, $sp, 16 +; LA64F-NEXT: move $a1, $s0 ; LA64F-NEXT: bl %plt(__adddf3) -; LA64F-NEXT: st.d $a0, $sp, 0 -; LA64F-NEXT: move $a0, $s0 +; LA64F-NEXT: st.d $a0, $sp, 8 +; LA64F-NEXT: move $a0, $s1 ; LA64F-NEXT: move $a1, $fp -; LA64F-NEXT: move $a2, $s1 -; LA64F-NEXT: move $a3, $s2 -; LA64F-NEXT: move $a4, $s3 -; LA64F-NEXT: move $a5, $s3 +; LA64F-NEXT: move $a2, $s2 +; LA64F-NEXT: move $a3, $s3 +; LA64F-NEXT: move $a4, $s4 +; LA64F-NEXT: move $a5, $s4 ; LA64F-NEXT: bl %plt(__atomic_compare_exchange) ; LA64F-NEXT: move $a1, $a0 -; LA64F-NEXT: ld.d $a0, $sp, 8 +; LA64F-NEXT: ld.d $a0, $sp, 16 ; LA64F-NEXT: beqz $a1, .LBB4_1 ; LA64F-NEXT: # %bb.2: # %atomicrmw.end -; LA64F-NEXT: ld.d $s3, $sp, 16 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s2, $sp, 24 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s1, $sp, 32 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s0, $sp, 40 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $fp, $sp, 48 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $ra, $sp, 56 # 8-byte Folded Reload -; LA64F-NEXT: addi.d $sp, $sp, 64 +; LA64F-NEXT: ld.d $s4, $sp, 24 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s3, $sp, 32 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s2, $sp, 40 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s1, $sp, 48 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s0, $sp, 56 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $fp, $sp, 64 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $ra, $sp, 72 # 8-byte Folded Reload +; LA64F-NEXT: addi.d $sp, $sp, 80 ; LA64F-NEXT: ret ; ; LA64D-LABEL: double_fadd_acquire: @@ -359,7 +362,8 @@ ; LA64D-NEXT: move $fp, $a0 ; LA64D-NEXT: fld.d $fa0, $a0, 0 ; LA64D-NEXT: addi.d $a0, $zero, 1 -; LA64D-NEXT: movgr2fr.d $fs0, $a0 +; LA64D-NEXT: movgr2fr.d $fa1, $a0 +; LA64D-NEXT: ffint.d.l $fs0, $fa1 ; LA64D-NEXT: ori $s0, $zero, 8 ; LA64D-NEXT: addi.d $s1, $sp, 16 ; LA64D-NEXT: addi.d $s2, $sp, 8 @@ -367,8 +371,7 @@ ; LA64D-NEXT: .LBB4_1: # %atomicrmw.start ; LA64D-NEXT: # =>This Inner Loop Header: Depth=1 ; LA64D-NEXT: fst.d $fa0, $sp, 16 -; LA64D-NEXT: ffint.d.l $fa1, $fs0 -; LA64D-NEXT: fadd.d $fa0, $fa0, $fa1 +; LA64D-NEXT: fadd.d $fa0, $fa0, $fs0 ; LA64D-NEXT: fst.d $fa0, $sp, 8 ; LA64D-NEXT: move $a0, $s0 ; LA64D-NEXT: move $a1, $fp @@ -396,43 +399,46 @@ define double @double_fsub_acquire(ptr %p) nounwind { ; LA64F-LABEL: double_fsub_acquire: ; LA64F: # %bb.0: -; LA64F-NEXT: addi.d $sp, $sp, -64 -; LA64F-NEXT: st.d $ra, $sp, 56 # 8-byte Folded Spill -; LA64F-NEXT: st.d $fp, $sp, 48 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s0, $sp, 40 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s1, $sp, 32 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s2, $sp, 24 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s3, $sp, 16 # 8-byte Folded Spill +; LA64F-NEXT: addi.d $sp, $sp, -80 +; LA64F-NEXT: st.d $ra, $sp, 72 # 8-byte Folded Spill +; LA64F-NEXT: st.d $fp, $sp, 64 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s0, $sp, 56 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s1, $sp, 48 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s2, $sp, 40 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s3, $sp, 32 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s4, $sp, 24 # 8-byte Folded Spill ; LA64F-NEXT: move $fp, $a0 ; LA64F-NEXT: ld.d $a0, $a0, 0 -; LA64F-NEXT: ori $s0, $zero, 8 -; LA64F-NEXT: addi.d $s1, $sp, 8 -; LA64F-NEXT: addi.d $s2, $sp, 0 -; LA64F-NEXT: ori $s3, $zero, 2 +; LA64F-NEXT: lu52i.d $s0, $zero, -1025 +; LA64F-NEXT: ori $s1, $zero, 8 +; LA64F-NEXT: addi.d $s2, $sp, 16 +; LA64F-NEXT: addi.d $s3, $sp, 8 +; LA64F-NEXT: ori $s4, $zero, 2 ; LA64F-NEXT: .LBB5_1: # %atomicrmw.start ; LA64F-NEXT: # =>This Inner Loop Header: Depth=1 -; LA64F-NEXT: st.d $a0, $sp, 8 -; LA64F-NEXT: lu52i.d $a1, $zero, -1025 +; LA64F-NEXT: st.d $a0, $sp, 16 +; LA64F-NEXT: move $a1, $s0 ; LA64F-NEXT: bl %plt(__adddf3) -; LA64F-NEXT: st.d $a0, $sp, 0 -; LA64F-NEXT: move $a0, $s0 +; LA64F-NEXT: st.d $a0, $sp, 8 +; LA64F-NEXT: move $a0, $s1 ; LA64F-NEXT: move $a1, $fp -; LA64F-NEXT: move $a2, $s1 -; LA64F-NEXT: move $a3, $s2 -; LA64F-NEXT: move $a4, $s3 -; LA64F-NEXT: move $a5, $s3 +; LA64F-NEXT: move $a2, $s2 +; LA64F-NEXT: move $a3, $s3 +; LA64F-NEXT: move $a4, $s4 +; LA64F-NEXT: move $a5, $s4 ; LA64F-NEXT: bl %plt(__atomic_compare_exchange) ; LA64F-NEXT: move $a1, $a0 -; LA64F-NEXT: ld.d $a0, $sp, 8 +; LA64F-NEXT: ld.d $a0, $sp, 16 ; LA64F-NEXT: beqz $a1, .LBB5_1 ; LA64F-NEXT: # %bb.2: # %atomicrmw.end -; LA64F-NEXT: ld.d $s3, $sp, 16 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s2, $sp, 24 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s1, $sp, 32 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s0, $sp, 40 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $fp, $sp, 48 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $ra, $sp, 56 # 8-byte Folded Reload -; LA64F-NEXT: addi.d $sp, $sp, 64 +; LA64F-NEXT: ld.d $s4, $sp, 24 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s3, $sp, 32 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s2, $sp, 40 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s1, $sp, 48 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s0, $sp, 56 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $fp, $sp, 64 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $ra, $sp, 72 # 8-byte Folded Reload +; LA64F-NEXT: addi.d $sp, $sp, 80 ; LA64F-NEXT: ret ; ; LA64D-LABEL: double_fsub_acquire: @@ -485,43 +491,46 @@ define double @double_fmin_acquire(ptr %p) nounwind { ; LA64F-LABEL: double_fmin_acquire: ; LA64F: # %bb.0: -; LA64F-NEXT: addi.d $sp, $sp, -64 -; LA64F-NEXT: st.d $ra, $sp, 56 # 8-byte Folded Spill -; LA64F-NEXT: st.d $fp, $sp, 48 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s0, $sp, 40 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s1, $sp, 32 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s2, $sp, 24 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s3, $sp, 16 # 8-byte Folded Spill +; LA64F-NEXT: addi.d $sp, $sp, -80 +; LA64F-NEXT: st.d $ra, $sp, 72 # 8-byte Folded Spill +; LA64F-NEXT: st.d $fp, $sp, 64 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s0, $sp, 56 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s1, $sp, 48 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s2, $sp, 40 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s3, $sp, 32 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s4, $sp, 24 # 8-byte Folded Spill ; LA64F-NEXT: move $fp, $a0 ; LA64F-NEXT: ld.d $a0, $a0, 0 -; LA64F-NEXT: ori $s0, $zero, 8 -; LA64F-NEXT: addi.d $s1, $sp, 8 -; LA64F-NEXT: addi.d $s2, $sp, 0 -; LA64F-NEXT: ori $s3, $zero, 2 +; LA64F-NEXT: lu52i.d $s0, $zero, 1023 +; LA64F-NEXT: ori $s1, $zero, 8 +; LA64F-NEXT: addi.d $s2, $sp, 16 +; LA64F-NEXT: addi.d $s3, $sp, 8 +; LA64F-NEXT: ori $s4, $zero, 2 ; LA64F-NEXT: .LBB6_1: # %atomicrmw.start ; LA64F-NEXT: # =>This Inner Loop Header: Depth=1 -; LA64F-NEXT: st.d $a0, $sp, 8 -; LA64F-NEXT: lu52i.d $a1, $zero, 1023 +; LA64F-NEXT: st.d $a0, $sp, 16 +; LA64F-NEXT: move $a1, $s0 ; LA64F-NEXT: bl %plt(fmin) -; LA64F-NEXT: st.d $a0, $sp, 0 -; LA64F-NEXT: move $a0, $s0 +; LA64F-NEXT: st.d $a0, $sp, 8 +; LA64F-NEXT: move $a0, $s1 ; LA64F-NEXT: move $a1, $fp -; LA64F-NEXT: move $a2, $s1 -; LA64F-NEXT: move $a3, $s2 -; LA64F-NEXT: move $a4, $s3 -; LA64F-NEXT: move $a5, $s3 +; LA64F-NEXT: move $a2, $s2 +; LA64F-NEXT: move $a3, $s3 +; LA64F-NEXT: move $a4, $s4 +; LA64F-NEXT: move $a5, $s4 ; LA64F-NEXT: bl %plt(__atomic_compare_exchange) ; LA64F-NEXT: move $a1, $a0 -; LA64F-NEXT: ld.d $a0, $sp, 8 +; LA64F-NEXT: ld.d $a0, $sp, 16 ; LA64F-NEXT: beqz $a1, .LBB6_1 ; LA64F-NEXT: # %bb.2: # %atomicrmw.end -; LA64F-NEXT: ld.d $s3, $sp, 16 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s2, $sp, 24 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s1, $sp, 32 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s0, $sp, 40 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $fp, $sp, 48 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $ra, $sp, 56 # 8-byte Folded Reload -; LA64F-NEXT: addi.d $sp, $sp, 64 +; LA64F-NEXT: ld.d $s4, $sp, 24 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s3, $sp, 32 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s2, $sp, 40 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s1, $sp, 48 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s0, $sp, 56 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $fp, $sp, 64 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $ra, $sp, 72 # 8-byte Folded Reload +; LA64F-NEXT: addi.d $sp, $sp, 80 ; LA64F-NEXT: ret ; ; LA64D-LABEL: double_fmin_acquire: @@ -537,7 +546,8 @@ ; LA64D-NEXT: move $fp, $a0 ; LA64D-NEXT: fld.d $fa0, $a0, 0 ; LA64D-NEXT: addi.d $a0, $zero, 1 -; LA64D-NEXT: movgr2fr.d $fs0, $a0 +; LA64D-NEXT: movgr2fr.d $fa1, $a0 +; LA64D-NEXT: ffint.d.l $fs0, $fa1 ; LA64D-NEXT: ori $s0, $zero, 8 ; LA64D-NEXT: addi.d $s1, $sp, 16 ; LA64D-NEXT: addi.d $s2, $sp, 8 @@ -546,8 +556,7 @@ ; LA64D-NEXT: # =>This Inner Loop Header: Depth=1 ; LA64D-NEXT: fst.d $fa0, $sp, 16 ; LA64D-NEXT: fmax.d $fa0, $fa0, $fa0 -; LA64D-NEXT: ffint.d.l $fa1, $fs0 -; LA64D-NEXT: fmin.d $fa0, $fa0, $fa1 +; LA64D-NEXT: fmin.d $fa0, $fa0, $fs0 ; LA64D-NEXT: fst.d $fa0, $sp, 8 ; LA64D-NEXT: move $a0, $s0 ; LA64D-NEXT: move $a1, $fp @@ -575,43 +584,46 @@ define double @double_fmax_acquire(ptr %p) nounwind { ; LA64F-LABEL: double_fmax_acquire: ; LA64F: # %bb.0: -; LA64F-NEXT: addi.d $sp, $sp, -64 -; LA64F-NEXT: st.d $ra, $sp, 56 # 8-byte Folded Spill -; LA64F-NEXT: st.d $fp, $sp, 48 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s0, $sp, 40 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s1, $sp, 32 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s2, $sp, 24 # 8-byte Folded Spill -; LA64F-NEXT: st.d $s3, $sp, 16 # 8-byte Folded Spill +; LA64F-NEXT: addi.d $sp, $sp, -80 +; LA64F-NEXT: st.d $ra, $sp, 72 # 8-byte Folded Spill +; LA64F-NEXT: st.d $fp, $sp, 64 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s0, $sp, 56 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s1, $sp, 48 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s2, $sp, 40 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s3, $sp, 32 # 8-byte Folded Spill +; LA64F-NEXT: st.d $s4, $sp, 24 # 8-byte Folded Spill ; LA64F-NEXT: move $fp, $a0 ; LA64F-NEXT: ld.d $a0, $a0, 0 -; LA64F-NEXT: ori $s0, $zero, 8 -; LA64F-NEXT: addi.d $s1, $sp, 8 -; LA64F-NEXT: addi.d $s2, $sp, 0 -; LA64F-NEXT: ori $s3, $zero, 2 +; LA64F-NEXT: lu52i.d $s0, $zero, 1023 +; LA64F-NEXT: ori $s1, $zero, 8 +; LA64F-NEXT: addi.d $s2, $sp, 16 +; LA64F-NEXT: addi.d $s3, $sp, 8 +; LA64F-NEXT: ori $s4, $zero, 2 ; LA64F-NEXT: .LBB7_1: # %atomicrmw.start ; LA64F-NEXT: # =>This Inner Loop Header: Depth=1 -; LA64F-NEXT: st.d $a0, $sp, 8 -; LA64F-NEXT: lu52i.d $a1, $zero, 1023 +; LA64F-NEXT: st.d $a0, $sp, 16 +; LA64F-NEXT: move $a1, $s0 ; LA64F-NEXT: bl %plt(fmax) -; LA64F-NEXT: st.d $a0, $sp, 0 -; LA64F-NEXT: move $a0, $s0 +; LA64F-NEXT: st.d $a0, $sp, 8 +; LA64F-NEXT: move $a0, $s1 ; LA64F-NEXT: move $a1, $fp -; LA64F-NEXT: move $a2, $s1 -; LA64F-NEXT: move $a3, $s2 -; LA64F-NEXT: move $a4, $s3 -; LA64F-NEXT: move $a5, $s3 +; LA64F-NEXT: move $a2, $s2 +; LA64F-NEXT: move $a3, $s3 +; LA64F-NEXT: move $a4, $s4 +; LA64F-NEXT: move $a5, $s4 ; LA64F-NEXT: bl %plt(__atomic_compare_exchange) ; LA64F-NEXT: move $a1, $a0 -; LA64F-NEXT: ld.d $a0, $sp, 8 +; LA64F-NEXT: ld.d $a0, $sp, 16 ; LA64F-NEXT: beqz $a1, .LBB7_1 ; LA64F-NEXT: # %bb.2: # %atomicrmw.end -; LA64F-NEXT: ld.d $s3, $sp, 16 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s2, $sp, 24 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s1, $sp, 32 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $s0, $sp, 40 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $fp, $sp, 48 # 8-byte Folded Reload -; LA64F-NEXT: ld.d $ra, $sp, 56 # 8-byte Folded Reload -; LA64F-NEXT: addi.d $sp, $sp, 64 +; LA64F-NEXT: ld.d $s4, $sp, 24 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s3, $sp, 32 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s2, $sp, 40 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s1, $sp, 48 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $s0, $sp, 56 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $fp, $sp, 64 # 8-byte Folded Reload +; LA64F-NEXT: ld.d $ra, $sp, 72 # 8-byte Folded Reload +; LA64F-NEXT: addi.d $sp, $sp, 80 ; LA64F-NEXT: ret ; ; LA64D-LABEL: double_fmax_acquire: @@ -627,7 +639,8 @@ ; LA64D-NEXT: move $fp, $a0 ; LA64D-NEXT: fld.d $fa0, $a0, 0 ; LA64D-NEXT: addi.d $a0, $zero, 1 -; LA64D-NEXT: movgr2fr.d $fs0, $a0 +; LA64D-NEXT: movgr2fr.d $fa1, $a0 +; LA64D-NEXT: ffint.d.l $fs0, $fa1 ; LA64D-NEXT: ori $s0, $zero, 8 ; LA64D-NEXT: addi.d $s1, $sp, 16 ; LA64D-NEXT: addi.d $s2, $sp, 8 @@ -636,8 +649,7 @@ ; LA64D-NEXT: # =>This Inner Loop Header: Depth=1 ; LA64D-NEXT: fst.d $fa0, $sp, 16 ; LA64D-NEXT: fmax.d $fa0, $fa0, $fa0 -; LA64D-NEXT: ffint.d.l $fa1, $fs0 -; LA64D-NEXT: fmax.d $fa0, $fa0, $fa1 +; LA64D-NEXT: fmax.d $fa0, $fa0, $fs0 ; LA64D-NEXT: fst.d $fa0, $sp, 8 ; LA64D-NEXT: move $a0, $s0 ; LA64D-NEXT: move $a1, $fp diff --git a/llvm/test/CodeGen/LoongArch/ir-instruction/fcmp-dbl.ll b/llvm/test/CodeGen/LoongArch/ir-instruction/fcmp-dbl.ll --- a/llvm/test/CodeGen/LoongArch/ir-instruction/fcmp-dbl.ll +++ b/llvm/test/CodeGen/LoongArch/ir-instruction/fcmp-dbl.ll @@ -268,8 +268,6 @@ ; LA32-NEXT: # %bb.1: # %if.then ; LA32-NEXT: ret ; LA32-NEXT: .LBB16_2: # %if.else -; LA32-NEXT: movgr2fr.w $fa1, $zero -; LA32-NEXT: movgr2frh.w $fa1, $zero ; LA32-NEXT: fcmp.clt.d $fcc0, $fa0, $fa1 ; LA32-NEXT: movcf2gr $a0, $fcc0 ; LA32-NEXT: ret @@ -308,9 +306,6 @@ ; LA32-NEXT: # %bb.1: # %if.then ; LA32-NEXT: ret ; LA32-NEXT: .LBB17_2: # %if.else -; LA32-NEXT: movgr2fr.w $fa1, $zero -; LA32-NEXT: movgr2frh.w $fa1, $zero -; LA32-NEXT: fcmp.ceq.d $fcc0, $fa0, $fa1 ; LA32-NEXT: movcf2gr $a0, $fcc0 ; LA32-NEXT: ret ; @@ -346,8 +341,6 @@ ; LA32-NEXT: # %bb.1: # %if.then ; LA32-NEXT: ret ; LA32-NEXT: .LBB18_2: # %if.else -; LA32-NEXT: movgr2fr.w $fa1, $zero -; LA32-NEXT: movgr2frh.w $fa1, $zero ; LA32-NEXT: fcmp.cle.d $fcc0, $fa0, $fa1 ; LA32-NEXT: movcf2gr $a0, $fcc0 ; LA32-NEXT: ret