diff --git a/llvm/lib/Target/RISCV/RISCVInstrInfoB.td b/llvm/lib/Target/RISCV/RISCVInstrInfoB.td --- a/llvm/lib/Target/RISCV/RISCVInstrInfoB.td +++ b/llvm/lib/Target/RISCV/RISCVInstrInfoB.td @@ -230,10 +230,10 @@ def SH3ADD : ALU_rr<0b0010000, 0b110, "sh3add">, Sched<[]>; } // Predicates = [HasStdExtZba] -let Predicates = [HasStdExtZbb] in { +let Predicates = [HasStdExtZbp] in { def SLO : ALU_rr<0b0010000, 0b001, "slo">, Sched<[]>; def SRO : ALU_rr<0b0010000, 0b101, "sro">, Sched<[]>; -} // Predicates = [HasStdExtZbb] +} // Predicates = [HasStdExtZbp] let Predicates = [HasStdExtZbbOrZbp] in { def ROL : ALU_rr<0b0110000, 0b001, "rol">, Sched<[]>; @@ -252,10 +252,10 @@ def GREV : ALU_rr<0b0110100, 0b101, "grev">, Sched<[]>; } // Predicates = [HasStdExtZbp] -let Predicates = [HasStdExtZbb] in { +let Predicates = [HasStdExtZbp] in { def SLOI : RVBShift_ri<0b00100, 0b001, OPC_OP_IMM, "sloi">, Sched<[]>; def SROI : RVBShift_ri<0b00100, 0b101, OPC_OP_IMM, "sroi">, Sched<[]>; -} // Predicates = [HasStdExtZbb] +} // Predicates = [HasStdExtZbp] let Predicates = [HasStdExtZbbOrZbp] in def RORI : RVBShift_ri<0b01100, 0b101, OPC_OP_IMM, "rori">, Sched<[]>; @@ -383,10 +383,10 @@ def SH3ADDUW : ALUW_rr<0b0010000, 0b110, "sh3add.uw">, Sched<[]>; } // Predicates = [HasStdExtZbb, IsRV64] -let Predicates = [HasStdExtZbb, IsRV64] in { +let Predicates = [HasStdExtZbp, IsRV64] in { def SLOW : ALUW_rr<0b0010000, 0b001, "slow">, Sched<[]>; def SROW : ALUW_rr<0b0010000, 0b101, "srow">, Sched<[]>; -} // Predicates = [HasStdExtZbb, IsRV64] +} // Predicates = [HasStdExtZbp, IsRV64] let Predicates = [HasStdExtZbbOrZbp, IsRV64] in { def ROLW : ALUW_rr<0b0110000, 0b001, "rolw">, Sched<[]>; @@ -405,10 +405,10 @@ def GREVW : ALUW_rr<0b0110100, 0b101, "grevw">, Sched<[]>; } // Predicates = [HasStdExtZbp, IsRV64] -let Predicates = [HasStdExtZbb, IsRV64] in { +let Predicates = [HasStdExtZbp, IsRV64] in { def SLOIW : RVBShiftW_ri<0b0010000, 0b001, OPC_OP_IMM_32, "sloiw">, Sched<[]>; def SROIW : RVBShiftW_ri<0b0010000, 0b101, OPC_OP_IMM_32, "sroiw">, Sched<[]>; -} // Predicates = [HasStdExtZbb, IsRV64] +} // Predicates = [HasStdExtZbp, IsRV64] let Predicates = [HasStdExtZbbOrZbp, IsRV64] in def RORIW : RVBShiftW_ri<0b0110000, 0b101, OPC_OP_IMM_32, "roriw">, Sched<[]>; @@ -637,12 +637,12 @@ def : Pat<(xor GPR:$rs1, (not GPR:$rs2)), (XNOR GPR:$rs1, GPR:$rs2)>; } // Predicates = [HasStdExtZbbOrZbp] -let Predicates = [HasStdExtZbb] in { +let Predicates = [HasStdExtZbp] in { def : Pat<(not (shl (not GPR:$rs1), GPR:$rs2)), (SLO GPR:$rs1, GPR:$rs2)>; def : Pat<(not (srl (not GPR:$rs1), GPR:$rs2)), (SRO GPR:$rs1, GPR:$rs2)>; -} // Predicates = [HasStdExtZbb] +} // Predicates = [HasStdExtZbp] let Predicates = [HasStdExtZbbOrZbp] in { def : Pat<(rotl GPR:$rs1, GPR:$rs2), (ROL GPR:$rs1, GPR:$rs2)>; @@ -671,12 +671,12 @@ (SBINVI GPR:$rs1, (SBSETINVXForm imm:$mask))>; } -let Predicates = [HasStdExtZbb] in { +let Predicates = [HasStdExtZbp] in { def : Pat<(SLOIPat GPR:$rs1, uimmlog2xlen:$shamt), (SLOI GPR:$rs1, uimmlog2xlen:$shamt)>; def : Pat<(SROIPat GPR:$rs1, uimmlog2xlen:$shamt), (SROI GPR:$rs1, uimmlog2xlen:$shamt)>; -} // Predicates = [HasStdExtZbb] +} // Predicates = [HasStdExtZbp] // There's no encoding for roli in the current version of the 'B' extension // (v0.92) as it can be implemented with rori by negating the immediate. @@ -856,12 +856,12 @@ (ADDUW GPR:$rs1, GPR:$rs2)>; } -let Predicates = [HasStdExtZbb, IsRV64] in { +let Predicates = [HasStdExtZbp, IsRV64] in { def : Pat<(not (riscv_sllw (not GPR:$rs1), GPR:$rs2)), (SLOW GPR:$rs1, GPR:$rs2)>; def : Pat<(not (riscv_srlw (not GPR:$rs1), GPR:$rs2)), (SROW GPR:$rs1, GPR:$rs2)>; -} // Predicates = [HasStdExtZbb, IsRV64] +} // Predicates = [HasStdExtZbp, IsRV64] let Predicates = [HasStdExtZbbOrZbp, IsRV64] in { def : Pat<(riscv_rolw GPR:$rs1, GPR:$rs2), @@ -902,12 +902,12 @@ } // Predicates = [HasStdExtZbs, IsRV64] -let Predicates = [HasStdExtZbb, IsRV64] in { +let Predicates = [HasStdExtZbp, IsRV64] in { def : Pat<(sext_inreg (SLOIPat GPR:$rs1, uimm5:$shamt), i32), (SLOIW GPR:$rs1, uimm5:$shamt)>; def : Pat<(SROIWPat GPR:$rs1, uimm5:$shamt), (SROIW GPR:$rs1, uimm5:$shamt)>; -} // Predicates = [HasStdExtZbb, IsRV64] +} // Predicates = [HasStdExtZbp, IsRV64] let Predicates = [HasStdExtZbp, IsRV64] in { def : Pat<(riscv_rorw (riscv_greviw GPR:$rs1, 24), (i64 16)), (GREVIW GPR:$rs1, 8)>; diff --git a/llvm/test/CodeGen/RISCV/rv32Zbb.ll b/llvm/test/CodeGen/RISCV/rv32Zbb.ll --- a/llvm/test/CodeGen/RISCV/rv32Zbb.ll +++ b/llvm/test/CodeGen/RISCV/rv32Zbb.ll @@ -6,308 +6,6 @@ ; RUN: llc -mtriple=riscv32 -mattr=+experimental-zbb -verify-machineinstrs < %s \ ; RUN: | FileCheck %s -check-prefix=RV32IBB -define i32 @slo_i32(i32 %a, i32 %b) nounwind { -; RV32I-LABEL: slo_i32: -; RV32I: # %bb.0: -; RV32I-NEXT: not a0, a0 -; RV32I-NEXT: sll a0, a0, a1 -; RV32I-NEXT: not a0, a0 -; RV32I-NEXT: ret -; -; RV32IB-LABEL: slo_i32: -; RV32IB: # %bb.0: -; RV32IB-NEXT: slo a0, a0, a1 -; RV32IB-NEXT: ret -; -; RV32IBB-LABEL: slo_i32: -; RV32IBB: # %bb.0: -; RV32IBB-NEXT: slo a0, a0, a1 -; RV32IBB-NEXT: ret - %neg = xor i32 %a, -1 - %shl = shl i32 %neg, %b - %neg1 = xor i32 %shl, -1 - ret i32 %neg1 -} - -; As we are not matching directly i64 code patterns on RV32 some i64 patterns -; don't have yet any matching bit manipulation instructions on RV32. -; This test is presented here in case future expansions of the experimental-b -; extension introduce instructions suitable for this pattern. - -define i64 @slo_i64(i64 %a, i64 %b) nounwind { -; RV32I-LABEL: slo_i64: -; RV32I: # %bb.0: -; RV32I-NEXT: addi a3, a2, -32 -; RV32I-NEXT: not a0, a0 -; RV32I-NEXT: bltz a3, .LBB1_2 -; RV32I-NEXT: # %bb.1: -; RV32I-NEXT: mv a2, zero -; RV32I-NEXT: sll a1, a0, a3 -; RV32I-NEXT: j .LBB1_3 -; RV32I-NEXT: .LBB1_2: -; RV32I-NEXT: not a1, a1 -; RV32I-NEXT: sll a1, a1, a2 -; RV32I-NEXT: addi a3, zero, 31 -; RV32I-NEXT: sub a3, a3, a2 -; RV32I-NEXT: srli a4, a0, 1 -; RV32I-NEXT: srl a3, a4, a3 -; RV32I-NEXT: or a1, a1, a3 -; RV32I-NEXT: sll a2, a0, a2 -; RV32I-NEXT: .LBB1_3: -; RV32I-NEXT: not a1, a1 -; RV32I-NEXT: not a0, a2 -; RV32I-NEXT: ret -; -; RV32IB-LABEL: slo_i64: -; RV32IB: # %bb.0: -; RV32IB-NEXT: not a0, a0 -; RV32IB-NEXT: not a1, a1 -; RV32IB-NEXT: sll a1, a1, a2 -; RV32IB-NEXT: addi a3, zero, 31 -; RV32IB-NEXT: sub a3, a3, a2 -; RV32IB-NEXT: srli a4, a0, 1 -; RV32IB-NEXT: srl a3, a4, a3 -; RV32IB-NEXT: or a1, a1, a3 -; RV32IB-NEXT: addi a3, a2, -32 -; RV32IB-NEXT: sll a4, a0, a3 -; RV32IB-NEXT: slti a5, a3, 0 -; RV32IB-NEXT: cmov a1, a5, a1, a4 -; RV32IB-NEXT: sll a0, a0, a2 -; RV32IB-NEXT: srai a2, a3, 31 -; RV32IB-NEXT: and a0, a2, a0 -; RV32IB-NEXT: not a1, a1 -; RV32IB-NEXT: not a0, a0 -; RV32IB-NEXT: ret -; -; RV32IBB-LABEL: slo_i64: -; RV32IBB: # %bb.0: -; RV32IBB-NEXT: addi a3, a2, -32 -; RV32IBB-NEXT: not a0, a0 -; RV32IBB-NEXT: bltz a3, .LBB1_2 -; RV32IBB-NEXT: # %bb.1: -; RV32IBB-NEXT: mv a2, zero -; RV32IBB-NEXT: sll a1, a0, a3 -; RV32IBB-NEXT: j .LBB1_3 -; RV32IBB-NEXT: .LBB1_2: -; RV32IBB-NEXT: not a1, a1 -; RV32IBB-NEXT: sll a1, a1, a2 -; RV32IBB-NEXT: addi a3, zero, 31 -; RV32IBB-NEXT: sub a3, a3, a2 -; RV32IBB-NEXT: srli a4, a0, 1 -; RV32IBB-NEXT: srl a3, a4, a3 -; RV32IBB-NEXT: or a1, a1, a3 -; RV32IBB-NEXT: sll a2, a0, a2 -; RV32IBB-NEXT: .LBB1_3: -; RV32IBB-NEXT: not a1, a1 -; RV32IBB-NEXT: not a0, a2 -; RV32IBB-NEXT: ret - %neg = xor i64 %a, -1 - %shl = shl i64 %neg, %b - %neg1 = xor i64 %shl, -1 - ret i64 %neg1 -} - -define i32 @sro_i32(i32 %a, i32 %b) nounwind { -; RV32I-LABEL: sro_i32: -; RV32I: # %bb.0: -; RV32I-NEXT: not a0, a0 -; RV32I-NEXT: srl a0, a0, a1 -; RV32I-NEXT: not a0, a0 -; RV32I-NEXT: ret -; -; RV32IB-LABEL: sro_i32: -; RV32IB: # %bb.0: -; RV32IB-NEXT: sro a0, a0, a1 -; RV32IB-NEXT: ret -; -; RV32IBB-LABEL: sro_i32: -; RV32IBB: # %bb.0: -; RV32IBB-NEXT: sro a0, a0, a1 -; RV32IBB-NEXT: ret - %neg = xor i32 %a, -1 - %shr = lshr i32 %neg, %b - %neg1 = xor i32 %shr, -1 - ret i32 %neg1 -} - -; As we are not matching directly i64 code patterns on RV32 some i64 patterns -; don't have yet any matching bit manipulation instructions on RV32. -; This test is presented here in case future expansions of the experimental-b -; extension introduce instructions suitable for this pattern. - -define i64 @sro_i64(i64 %a, i64 %b) nounwind { -; RV32I-LABEL: sro_i64: -; RV32I: # %bb.0: -; RV32I-NEXT: addi a3, a2, -32 -; RV32I-NEXT: not a1, a1 -; RV32I-NEXT: bltz a3, .LBB3_2 -; RV32I-NEXT: # %bb.1: -; RV32I-NEXT: mv a2, zero -; RV32I-NEXT: srl a0, a1, a3 -; RV32I-NEXT: j .LBB3_3 -; RV32I-NEXT: .LBB3_2: -; RV32I-NEXT: not a0, a0 -; RV32I-NEXT: srl a0, a0, a2 -; RV32I-NEXT: addi a3, zero, 31 -; RV32I-NEXT: sub a3, a3, a2 -; RV32I-NEXT: slli a4, a1, 1 -; RV32I-NEXT: sll a3, a4, a3 -; RV32I-NEXT: or a0, a0, a3 -; RV32I-NEXT: srl a2, a1, a2 -; RV32I-NEXT: .LBB3_3: -; RV32I-NEXT: not a0, a0 -; RV32I-NEXT: not a1, a2 -; RV32I-NEXT: ret -; -; RV32IB-LABEL: sro_i64: -; RV32IB: # %bb.0: -; RV32IB-NEXT: not a1, a1 -; RV32IB-NEXT: not a0, a0 -; RV32IB-NEXT: srl a0, a0, a2 -; RV32IB-NEXT: addi a3, zero, 31 -; RV32IB-NEXT: sub a3, a3, a2 -; RV32IB-NEXT: slli a4, a1, 1 -; RV32IB-NEXT: sll a3, a4, a3 -; RV32IB-NEXT: or a0, a0, a3 -; RV32IB-NEXT: addi a3, a2, -32 -; RV32IB-NEXT: srl a4, a1, a3 -; RV32IB-NEXT: slti a5, a3, 0 -; RV32IB-NEXT: cmov a0, a5, a0, a4 -; RV32IB-NEXT: srl a1, a1, a2 -; RV32IB-NEXT: srai a2, a3, 31 -; RV32IB-NEXT: and a1, a2, a1 -; RV32IB-NEXT: not a0, a0 -; RV32IB-NEXT: not a1, a1 -; RV32IB-NEXT: ret -; -; RV32IBB-LABEL: sro_i64: -; RV32IBB: # %bb.0: -; RV32IBB-NEXT: addi a3, a2, -32 -; RV32IBB-NEXT: not a1, a1 -; RV32IBB-NEXT: bltz a3, .LBB3_2 -; RV32IBB-NEXT: # %bb.1: -; RV32IBB-NEXT: mv a2, zero -; RV32IBB-NEXT: srl a0, a1, a3 -; RV32IBB-NEXT: j .LBB3_3 -; RV32IBB-NEXT: .LBB3_2: -; RV32IBB-NEXT: not a0, a0 -; RV32IBB-NEXT: srl a0, a0, a2 -; RV32IBB-NEXT: addi a3, zero, 31 -; RV32IBB-NEXT: sub a3, a3, a2 -; RV32IBB-NEXT: slli a4, a1, 1 -; RV32IBB-NEXT: sll a3, a4, a3 -; RV32IBB-NEXT: or a0, a0, a3 -; RV32IBB-NEXT: srl a2, a1, a2 -; RV32IBB-NEXT: .LBB3_3: -; RV32IBB-NEXT: not a0, a0 -; RV32IBB-NEXT: not a1, a2 -; RV32IBB-NEXT: ret - %neg = xor i64 %a, -1 - %shr = lshr i64 %neg, %b - %neg1 = xor i64 %shr, -1 - ret i64 %neg1 -} - -define i32 @sloi_i32(i32 %a) nounwind { -; RV32I-LABEL: sloi_i32: -; RV32I: # %bb.0: -; RV32I-NEXT: slli a0, a0, 1 -; RV32I-NEXT: ori a0, a0, 1 -; RV32I-NEXT: ret -; -; RV32IB-LABEL: sloi_i32: -; RV32IB: # %bb.0: -; RV32IB-NEXT: sloi a0, a0, 1 -; RV32IB-NEXT: ret -; -; RV32IBB-LABEL: sloi_i32: -; RV32IBB: # %bb.0: -; RV32IBB-NEXT: sloi a0, a0, 1 -; RV32IBB-NEXT: ret - %neg = shl i32 %a, 1 - %neg12 = or i32 %neg, 1 - ret i32 %neg12 -} - -define i64 @sloi_i64(i64 %a) nounwind { -; RV32I-LABEL: sloi_i64: -; RV32I: # %bb.0: -; RV32I-NEXT: srli a2, a0, 31 -; RV32I-NEXT: slli a1, a1, 1 -; RV32I-NEXT: or a1, a1, a2 -; RV32I-NEXT: slli a0, a0, 1 -; RV32I-NEXT: ori a0, a0, 1 -; RV32I-NEXT: ret -; -; RV32IB-LABEL: sloi_i64: -; RV32IB: # %bb.0: -; RV32IB-NEXT: fsri a1, a0, a1, 31 -; RV32IB-NEXT: sloi a0, a0, 1 -; RV32IB-NEXT: ret -; -; RV32IBB-LABEL: sloi_i64: -; RV32IBB: # %bb.0: -; RV32IBB-NEXT: srli a2, a0, 31 -; RV32IBB-NEXT: slli a1, a1, 1 -; RV32IBB-NEXT: or a1, a1, a2 -; RV32IBB-NEXT: sloi a0, a0, 1 -; RV32IBB-NEXT: ret - %neg = shl i64 %a, 1 - %neg12 = or i64 %neg, 1 - ret i64 %neg12 -} - -define i32 @sroi_i32(i32 %a) nounwind { -; RV32I-LABEL: sroi_i32: -; RV32I: # %bb.0: -; RV32I-NEXT: srli a0, a0, 1 -; RV32I-NEXT: lui a1, 524288 -; RV32I-NEXT: or a0, a0, a1 -; RV32I-NEXT: ret -; -; RV32IB-LABEL: sroi_i32: -; RV32IB: # %bb.0: -; RV32IB-NEXT: sroi a0, a0, 1 -; RV32IB-NEXT: ret -; -; RV32IBB-LABEL: sroi_i32: -; RV32IBB: # %bb.0: -; RV32IBB-NEXT: sroi a0, a0, 1 -; RV32IBB-NEXT: ret - %neg = lshr i32 %a, 1 - %neg12 = or i32 %neg, -2147483648 - ret i32 %neg12 -} - -define i64 @sroi_i64(i64 %a) nounwind { -; RV32I-LABEL: sroi_i64: -; RV32I: # %bb.0: -; RV32I-NEXT: slli a2, a1, 31 -; RV32I-NEXT: srli a0, a0, 1 -; RV32I-NEXT: or a0, a0, a2 -; RV32I-NEXT: srli a1, a1, 1 -; RV32I-NEXT: lui a2, 524288 -; RV32I-NEXT: or a1, a1, a2 -; RV32I-NEXT: ret -; -; RV32IB-LABEL: sroi_i64: -; RV32IB: # %bb.0: -; RV32IB-NEXT: fsri a0, a0, a1, 1 -; RV32IB-NEXT: sroi a1, a1, 1 -; RV32IB-NEXT: ret -; -; RV32IBB-LABEL: sroi_i64: -; RV32IBB: # %bb.0: -; RV32IBB-NEXT: slli a2, a1, 31 -; RV32IBB-NEXT: srli a0, a0, 1 -; RV32IBB-NEXT: or a0, a0, a2 -; RV32IBB-NEXT: sroi a1, a1, 1 -; RV32IBB-NEXT: ret - %neg = lshr i64 %a, 1 - %neg12 = or i64 %neg, -9223372036854775808 - ret i64 %neg12 -} - declare i32 @llvm.ctlz.i32(i32, i1) define i32 @ctlz_i32(i32 %a) nounwind { @@ -315,7 +13,7 @@ ; RV32I: # %bb.0: ; RV32I-NEXT: addi sp, sp, -16 ; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill -; RV32I-NEXT: beqz a0, .LBB8_2 +; RV32I-NEXT: beqz a0, .LBB0_2 ; RV32I-NEXT: # %bb.1: # %cond.false ; RV32I-NEXT: srli a1, a0, 1 ; RV32I-NEXT: or a0, a0, a1 @@ -348,10 +46,10 @@ ; RV32I-NEXT: addi a1, a1, 257 ; RV32I-NEXT: call __mulsi3@plt ; RV32I-NEXT: srli a0, a0, 24 -; RV32I-NEXT: j .LBB8_3 -; RV32I-NEXT: .LBB8_2: +; RV32I-NEXT: j .LBB0_3 +; RV32I-NEXT: .LBB0_2: ; RV32I-NEXT: addi a0, zero, 32 -; RV32I-NEXT: .LBB8_3: # %cond.end +; RV32I-NEXT: .LBB0_3: # %cond.end ; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload ; RV32I-NEXT: addi sp, sp, 16 ; RV32I-NEXT: ret @@ -440,14 +138,14 @@ ; RV32I-NEXT: and a0, a0, s6 ; RV32I-NEXT: mv a1, s0 ; RV32I-NEXT: call __mulsi3@plt -; RV32I-NEXT: bnez s3, .LBB9_2 +; RV32I-NEXT: bnez s3, .LBB1_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: srli a0, a0, 24 ; RV32I-NEXT: addi a0, a0, 32 -; RV32I-NEXT: j .LBB9_3 -; RV32I-NEXT: .LBB9_2: +; RV32I-NEXT: j .LBB1_3 +; RV32I-NEXT: .LBB1_2: ; RV32I-NEXT: srli a0, s2, 24 -; RV32I-NEXT: .LBB9_3: +; RV32I-NEXT: .LBB1_3: ; RV32I-NEXT: mv a1, zero ; RV32I-NEXT: lw s6, 0(sp) # 4-byte Folded Reload ; RV32I-NEXT: lw s5, 4(sp) # 4-byte Folded Reload @@ -471,13 +169,13 @@ ; ; RV32IBB-LABEL: ctlz_i64: ; RV32IBB: # %bb.0: -; RV32IBB-NEXT: bnez a1, .LBB9_2 +; RV32IBB-NEXT: bnez a1, .LBB1_2 ; RV32IBB-NEXT: # %bb.1: ; RV32IBB-NEXT: clz a0, a0 ; RV32IBB-NEXT: addi a0, a0, 32 ; RV32IBB-NEXT: mv a1, zero ; RV32IBB-NEXT: ret -; RV32IBB-NEXT: .LBB9_2: +; RV32IBB-NEXT: .LBB1_2: ; RV32IBB-NEXT: clz a0, a1 ; RV32IBB-NEXT: mv a1, zero ; RV32IBB-NEXT: ret @@ -492,7 +190,7 @@ ; RV32I: # %bb.0: ; RV32I-NEXT: addi sp, sp, -16 ; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill -; RV32I-NEXT: beqz a0, .LBB10_2 +; RV32I-NEXT: beqz a0, .LBB2_2 ; RV32I-NEXT: # %bb.1: # %cond.false ; RV32I-NEXT: addi a1, a0, -1 ; RV32I-NEXT: not a0, a0 @@ -517,10 +215,10 @@ ; RV32I-NEXT: addi a1, a1, 257 ; RV32I-NEXT: call __mulsi3@plt ; RV32I-NEXT: srli a0, a0, 24 -; RV32I-NEXT: j .LBB10_3 -; RV32I-NEXT: .LBB10_2: +; RV32I-NEXT: j .LBB2_3 +; RV32I-NEXT: .LBB2_2: ; RV32I-NEXT: addi a0, zero, 32 -; RV32I-NEXT: .LBB10_3: # %cond.end +; RV32I-NEXT: .LBB2_3: # %cond.end ; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload ; RV32I-NEXT: addi sp, sp, 16 ; RV32I-NEXT: ret @@ -593,14 +291,14 @@ ; RV32I-NEXT: and a0, a0, s6 ; RV32I-NEXT: mv a1, s1 ; RV32I-NEXT: call __mulsi3@plt -; RV32I-NEXT: bnez s4, .LBB11_2 +; RV32I-NEXT: bnez s4, .LBB3_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: srli a0, a0, 24 ; RV32I-NEXT: addi a0, a0, 32 -; RV32I-NEXT: j .LBB11_3 -; RV32I-NEXT: .LBB11_2: +; RV32I-NEXT: j .LBB3_3 +; RV32I-NEXT: .LBB3_2: ; RV32I-NEXT: srli a0, s2, 24 -; RV32I-NEXT: .LBB11_3: +; RV32I-NEXT: .LBB3_3: ; RV32I-NEXT: mv a1, zero ; RV32I-NEXT: lw s6, 0(sp) # 4-byte Folded Reload ; RV32I-NEXT: lw s5, 4(sp) # 4-byte Folded Reload @@ -624,13 +322,13 @@ ; ; RV32IBB-LABEL: cttz_i64: ; RV32IBB: # %bb.0: -; RV32IBB-NEXT: bnez a0, .LBB11_2 +; RV32IBB-NEXT: bnez a0, .LBB3_2 ; RV32IBB-NEXT: # %bb.1: ; RV32IBB-NEXT: ctz a0, a1 ; RV32IBB-NEXT: addi a0, a0, 32 ; RV32IBB-NEXT: mv a1, zero ; RV32IBB-NEXT: ret -; RV32IBB-NEXT: .LBB11_2: +; RV32IBB-NEXT: .LBB3_2: ; RV32IBB-NEXT: ctz a0, a0 ; RV32IBB-NEXT: mv a1, zero ; RV32IBB-NEXT: ret @@ -854,10 +552,10 @@ define i32 @min_i32(i32 %a, i32 %b) nounwind { ; RV32I-LABEL: min_i32: ; RV32I: # %bb.0: -; RV32I-NEXT: blt a0, a1, .LBB18_2 +; RV32I-NEXT: blt a0, a1, .LBB10_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: mv a0, a1 -; RV32I-NEXT: .LBB18_2: +; RV32I-NEXT: .LBB10_2: ; RV32I-NEXT: ret ; ; RV32IB-LABEL: min_i32: @@ -882,18 +580,18 @@ define i64 @min_i64(i64 %a, i64 %b) nounwind { ; RV32I-LABEL: min_i64: ; RV32I: # %bb.0: -; RV32I-NEXT: beq a1, a3, .LBB19_2 +; RV32I-NEXT: beq a1, a3, .LBB11_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: slt a4, a1, a3 -; RV32I-NEXT: beqz a4, .LBB19_3 -; RV32I-NEXT: j .LBB19_4 -; RV32I-NEXT: .LBB19_2: +; RV32I-NEXT: beqz a4, .LBB11_3 +; RV32I-NEXT: j .LBB11_4 +; RV32I-NEXT: .LBB11_2: ; RV32I-NEXT: sltu a4, a0, a2 -; RV32I-NEXT: bnez a4, .LBB19_4 -; RV32I-NEXT: .LBB19_3: +; RV32I-NEXT: bnez a4, .LBB11_4 +; RV32I-NEXT: .LBB11_3: ; RV32I-NEXT: mv a0, a2 ; RV32I-NEXT: mv a1, a3 -; RV32I-NEXT: .LBB19_4: +; RV32I-NEXT: .LBB11_4: ; RV32I-NEXT: ret ; ; RV32IB-LABEL: min_i64: @@ -909,16 +607,16 @@ ; RV32IBB-LABEL: min_i64: ; RV32IBB: # %bb.0: ; RV32IBB-NEXT: mv a4, a0 -; RV32IBB-NEXT: bge a1, a3, .LBB19_3 +; RV32IBB-NEXT: bge a1, a3, .LBB11_3 ; RV32IBB-NEXT: # %bb.1: -; RV32IBB-NEXT: beq a1, a3, .LBB19_4 -; RV32IBB-NEXT: .LBB19_2: +; RV32IBB-NEXT: beq a1, a3, .LBB11_4 +; RV32IBB-NEXT: .LBB11_2: ; RV32IBB-NEXT: min a1, a1, a3 ; RV32IBB-NEXT: ret -; RV32IBB-NEXT: .LBB19_3: +; RV32IBB-NEXT: .LBB11_3: ; RV32IBB-NEXT: mv a0, a2 -; RV32IBB-NEXT: bne a1, a3, .LBB19_2 -; RV32IBB-NEXT: .LBB19_4: +; RV32IBB-NEXT: bne a1, a3, .LBB11_2 +; RV32IBB-NEXT: .LBB11_4: ; RV32IBB-NEXT: minu a0, a4, a2 ; RV32IBB-NEXT: min a1, a1, a3 ; RV32IBB-NEXT: ret @@ -930,10 +628,10 @@ define i32 @max_i32(i32 %a, i32 %b) nounwind { ; RV32I-LABEL: max_i32: ; RV32I: # %bb.0: -; RV32I-NEXT: blt a1, a0, .LBB20_2 +; RV32I-NEXT: blt a1, a0, .LBB12_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: mv a0, a1 -; RV32I-NEXT: .LBB20_2: +; RV32I-NEXT: .LBB12_2: ; RV32I-NEXT: ret ; ; RV32IB-LABEL: max_i32: @@ -958,18 +656,18 @@ define i64 @max_i64(i64 %a, i64 %b) nounwind { ; RV32I-LABEL: max_i64: ; RV32I: # %bb.0: -; RV32I-NEXT: beq a1, a3, .LBB21_2 +; RV32I-NEXT: beq a1, a3, .LBB13_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: slt a4, a3, a1 -; RV32I-NEXT: beqz a4, .LBB21_3 -; RV32I-NEXT: j .LBB21_4 -; RV32I-NEXT: .LBB21_2: +; RV32I-NEXT: beqz a4, .LBB13_3 +; RV32I-NEXT: j .LBB13_4 +; RV32I-NEXT: .LBB13_2: ; RV32I-NEXT: sltu a4, a2, a0 -; RV32I-NEXT: bnez a4, .LBB21_4 -; RV32I-NEXT: .LBB21_3: +; RV32I-NEXT: bnez a4, .LBB13_4 +; RV32I-NEXT: .LBB13_3: ; RV32I-NEXT: mv a0, a2 ; RV32I-NEXT: mv a1, a3 -; RV32I-NEXT: .LBB21_4: +; RV32I-NEXT: .LBB13_4: ; RV32I-NEXT: ret ; ; RV32IB-LABEL: max_i64: @@ -985,16 +683,16 @@ ; RV32IBB-LABEL: max_i64: ; RV32IBB: # %bb.0: ; RV32IBB-NEXT: mv a4, a0 -; RV32IBB-NEXT: bge a3, a1, .LBB21_3 +; RV32IBB-NEXT: bge a3, a1, .LBB13_3 ; RV32IBB-NEXT: # %bb.1: -; RV32IBB-NEXT: beq a1, a3, .LBB21_4 -; RV32IBB-NEXT: .LBB21_2: +; RV32IBB-NEXT: beq a1, a3, .LBB13_4 +; RV32IBB-NEXT: .LBB13_2: ; RV32IBB-NEXT: max a1, a1, a3 ; RV32IBB-NEXT: ret -; RV32IBB-NEXT: .LBB21_3: +; RV32IBB-NEXT: .LBB13_3: ; RV32IBB-NEXT: mv a0, a2 -; RV32IBB-NEXT: bne a1, a3, .LBB21_2 -; RV32IBB-NEXT: .LBB21_4: +; RV32IBB-NEXT: bne a1, a3, .LBB13_2 +; RV32IBB-NEXT: .LBB13_4: ; RV32IBB-NEXT: maxu a0, a4, a2 ; RV32IBB-NEXT: max a1, a1, a3 ; RV32IBB-NEXT: ret @@ -1006,10 +704,10 @@ define i32 @minu_i32(i32 %a, i32 %b) nounwind { ; RV32I-LABEL: minu_i32: ; RV32I: # %bb.0: -; RV32I-NEXT: bltu a0, a1, .LBB22_2 +; RV32I-NEXT: bltu a0, a1, .LBB14_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: mv a0, a1 -; RV32I-NEXT: .LBB22_2: +; RV32I-NEXT: .LBB14_2: ; RV32I-NEXT: ret ; ; RV32IB-LABEL: minu_i32: @@ -1034,18 +732,18 @@ define i64 @minu_i64(i64 %a, i64 %b) nounwind { ; RV32I-LABEL: minu_i64: ; RV32I: # %bb.0: -; RV32I-NEXT: beq a1, a3, .LBB23_2 +; RV32I-NEXT: beq a1, a3, .LBB15_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: sltu a4, a1, a3 -; RV32I-NEXT: beqz a4, .LBB23_3 -; RV32I-NEXT: j .LBB23_4 -; RV32I-NEXT: .LBB23_2: +; RV32I-NEXT: beqz a4, .LBB15_3 +; RV32I-NEXT: j .LBB15_4 +; RV32I-NEXT: .LBB15_2: ; RV32I-NEXT: sltu a4, a0, a2 -; RV32I-NEXT: bnez a4, .LBB23_4 -; RV32I-NEXT: .LBB23_3: +; RV32I-NEXT: bnez a4, .LBB15_4 +; RV32I-NEXT: .LBB15_3: ; RV32I-NEXT: mv a0, a2 ; RV32I-NEXT: mv a1, a3 -; RV32I-NEXT: .LBB23_4: +; RV32I-NEXT: .LBB15_4: ; RV32I-NEXT: ret ; ; RV32IB-LABEL: minu_i64: @@ -1061,16 +759,16 @@ ; RV32IBB-LABEL: minu_i64: ; RV32IBB: # %bb.0: ; RV32IBB-NEXT: mv a4, a0 -; RV32IBB-NEXT: bgeu a1, a3, .LBB23_3 +; RV32IBB-NEXT: bgeu a1, a3, .LBB15_3 ; RV32IBB-NEXT: # %bb.1: -; RV32IBB-NEXT: beq a1, a3, .LBB23_4 -; RV32IBB-NEXT: .LBB23_2: +; RV32IBB-NEXT: beq a1, a3, .LBB15_4 +; RV32IBB-NEXT: .LBB15_2: ; RV32IBB-NEXT: minu a1, a1, a3 ; RV32IBB-NEXT: ret -; RV32IBB-NEXT: .LBB23_3: +; RV32IBB-NEXT: .LBB15_3: ; RV32IBB-NEXT: mv a0, a2 -; RV32IBB-NEXT: bne a1, a3, .LBB23_2 -; RV32IBB-NEXT: .LBB23_4: +; RV32IBB-NEXT: bne a1, a3, .LBB15_2 +; RV32IBB-NEXT: .LBB15_4: ; RV32IBB-NEXT: minu a0, a4, a2 ; RV32IBB-NEXT: minu a1, a1, a3 ; RV32IBB-NEXT: ret @@ -1082,10 +780,10 @@ define i32 @maxu_i32(i32 %a, i32 %b) nounwind { ; RV32I-LABEL: maxu_i32: ; RV32I: # %bb.0: -; RV32I-NEXT: bltu a1, a0, .LBB24_2 +; RV32I-NEXT: bltu a1, a0, .LBB16_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: mv a0, a1 -; RV32I-NEXT: .LBB24_2: +; RV32I-NEXT: .LBB16_2: ; RV32I-NEXT: ret ; ; RV32IB-LABEL: maxu_i32: @@ -1110,18 +808,18 @@ define i64 @maxu_i64(i64 %a, i64 %b) nounwind { ; RV32I-LABEL: maxu_i64: ; RV32I: # %bb.0: -; RV32I-NEXT: beq a1, a3, .LBB25_2 +; RV32I-NEXT: beq a1, a3, .LBB17_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: sltu a4, a3, a1 -; RV32I-NEXT: beqz a4, .LBB25_3 -; RV32I-NEXT: j .LBB25_4 -; RV32I-NEXT: .LBB25_2: +; RV32I-NEXT: beqz a4, .LBB17_3 +; RV32I-NEXT: j .LBB17_4 +; RV32I-NEXT: .LBB17_2: ; RV32I-NEXT: sltu a4, a2, a0 -; RV32I-NEXT: bnez a4, .LBB25_4 -; RV32I-NEXT: .LBB25_3: +; RV32I-NEXT: bnez a4, .LBB17_4 +; RV32I-NEXT: .LBB17_3: ; RV32I-NEXT: mv a0, a2 ; RV32I-NEXT: mv a1, a3 -; RV32I-NEXT: .LBB25_4: +; RV32I-NEXT: .LBB17_4: ; RV32I-NEXT: ret ; ; RV32IB-LABEL: maxu_i64: @@ -1137,16 +835,16 @@ ; RV32IBB-LABEL: maxu_i64: ; RV32IBB: # %bb.0: ; RV32IBB-NEXT: mv a4, a0 -; RV32IBB-NEXT: bgeu a3, a1, .LBB25_3 +; RV32IBB-NEXT: bgeu a3, a1, .LBB17_3 ; RV32IBB-NEXT: # %bb.1: -; RV32IBB-NEXT: beq a1, a3, .LBB25_4 -; RV32IBB-NEXT: .LBB25_2: +; RV32IBB-NEXT: beq a1, a3, .LBB17_4 +; RV32IBB-NEXT: .LBB17_2: ; RV32IBB-NEXT: maxu a1, a1, a3 ; RV32IBB-NEXT: ret -; RV32IBB-NEXT: .LBB25_3: +; RV32IBB-NEXT: .LBB17_3: ; RV32IBB-NEXT: mv a0, a2 -; RV32IBB-NEXT: bne a1, a3, .LBB25_2 -; RV32IBB-NEXT: .LBB25_4: +; RV32IBB-NEXT: bne a1, a3, .LBB17_2 +; RV32IBB-NEXT: .LBB17_4: ; RV32IBB-NEXT: maxu a0, a4, a2 ; RV32IBB-NEXT: maxu a1, a1, a3 ; RV32IBB-NEXT: ret @@ -1185,13 +883,13 @@ define i64 @abs_i64(i64 %x) { ; RV32I-LABEL: abs_i64: ; RV32I: # %bb.0: -; RV32I-NEXT: bgez a1, .LBB27_2 +; RV32I-NEXT: bgez a1, .LBB19_2 ; RV32I-NEXT: # %bb.1: ; RV32I-NEXT: snez a2, a0 ; RV32I-NEXT: neg a0, a0 ; RV32I-NEXT: add a1, a1, a2 ; RV32I-NEXT: neg a1, a1 -; RV32I-NEXT: .LBB27_2: +; RV32I-NEXT: .LBB19_2: ; RV32I-NEXT: ret ; ; RV32IB-LABEL: abs_i64: @@ -1208,13 +906,13 @@ ; ; RV32IBB-LABEL: abs_i64: ; RV32IBB: # %bb.0: -; RV32IBB-NEXT: bgez a1, .LBB27_2 +; RV32IBB-NEXT: bgez a1, .LBB19_2 ; RV32IBB-NEXT: # %bb.1: ; RV32IBB-NEXT: snez a2, a0 ; RV32IBB-NEXT: neg a0, a0 ; RV32IBB-NEXT: add a1, a1, a2 ; RV32IBB-NEXT: neg a1, a1 -; RV32IBB-NEXT: .LBB27_2: +; RV32IBB-NEXT: .LBB19_2: ; RV32IBB-NEXT: ret %abs = tail call i64 @llvm.abs.i64(i64 %x, i1 true) ret i64 %abs diff --git a/llvm/test/CodeGen/RISCV/rv32Zbp.ll b/llvm/test/CodeGen/RISCV/rv32Zbp.ll --- a/llvm/test/CodeGen/RISCV/rv32Zbp.ll +++ b/llvm/test/CodeGen/RISCV/rv32Zbp.ll @@ -6,6 +6,308 @@ ; RUN: llc -mtriple=riscv32 -mattr=+experimental-zbp -verify-machineinstrs < %s \ ; RUN: | FileCheck %s -check-prefix=RV32IBP +define i32 @slo_i32(i32 %a, i32 %b) nounwind { +; RV32I-LABEL: slo_i32: +; RV32I: # %bb.0: +; RV32I-NEXT: not a0, a0 +; RV32I-NEXT: sll a0, a0, a1 +; RV32I-NEXT: not a0, a0 +; RV32I-NEXT: ret +; +; RV32IB-LABEL: slo_i32: +; RV32IB: # %bb.0: +; RV32IB-NEXT: slo a0, a0, a1 +; RV32IB-NEXT: ret +; +; RV32IBP-LABEL: slo_i32: +; RV32IBP: # %bb.0: +; RV32IBP-NEXT: slo a0, a0, a1 +; RV32IBP-NEXT: ret + %neg = xor i32 %a, -1 + %shl = shl i32 %neg, %b + %neg1 = xor i32 %shl, -1 + ret i32 %neg1 +} + +; As we are not matching directly i64 code patterns on RV32 some i64 patterns +; don't have yet any matching bit manipulation instructions on RV32. +; This test is presented here in case future expansions of the experimental-b +; extension introduce instructions suitable for this pattern. + +define i64 @slo_i64(i64 %a, i64 %b) nounwind { +; RV32I-LABEL: slo_i64: +; RV32I: # %bb.0: +; RV32I-NEXT: addi a3, a2, -32 +; RV32I-NEXT: not a0, a0 +; RV32I-NEXT: bltz a3, .LBB1_2 +; RV32I-NEXT: # %bb.1: +; RV32I-NEXT: mv a2, zero +; RV32I-NEXT: sll a1, a0, a3 +; RV32I-NEXT: j .LBB1_3 +; RV32I-NEXT: .LBB1_2: +; RV32I-NEXT: not a1, a1 +; RV32I-NEXT: sll a1, a1, a2 +; RV32I-NEXT: addi a3, zero, 31 +; RV32I-NEXT: sub a3, a3, a2 +; RV32I-NEXT: srli a4, a0, 1 +; RV32I-NEXT: srl a3, a4, a3 +; RV32I-NEXT: or a1, a1, a3 +; RV32I-NEXT: sll a2, a0, a2 +; RV32I-NEXT: .LBB1_3: +; RV32I-NEXT: not a1, a1 +; RV32I-NEXT: not a0, a2 +; RV32I-NEXT: ret +; +; RV32IB-LABEL: slo_i64: +; RV32IB: # %bb.0: +; RV32IB-NEXT: not a0, a0 +; RV32IB-NEXT: not a1, a1 +; RV32IB-NEXT: sll a1, a1, a2 +; RV32IB-NEXT: addi a3, zero, 31 +; RV32IB-NEXT: sub a3, a3, a2 +; RV32IB-NEXT: srli a4, a0, 1 +; RV32IB-NEXT: srl a3, a4, a3 +; RV32IB-NEXT: or a1, a1, a3 +; RV32IB-NEXT: addi a3, a2, -32 +; RV32IB-NEXT: sll a4, a0, a3 +; RV32IB-NEXT: slti a5, a3, 0 +; RV32IB-NEXT: cmov a1, a5, a1, a4 +; RV32IB-NEXT: sll a0, a0, a2 +; RV32IB-NEXT: srai a2, a3, 31 +; RV32IB-NEXT: and a0, a2, a0 +; RV32IB-NEXT: not a1, a1 +; RV32IB-NEXT: not a0, a0 +; RV32IB-NEXT: ret +; +; RV32IBP-LABEL: slo_i64: +; RV32IBP: # %bb.0: +; RV32IBP-NEXT: addi a3, a2, -32 +; RV32IBP-NEXT: not a0, a0 +; RV32IBP-NEXT: bltz a3, .LBB1_2 +; RV32IBP-NEXT: # %bb.1: +; RV32IBP-NEXT: mv a2, zero +; RV32IBP-NEXT: sll a1, a0, a3 +; RV32IBP-NEXT: j .LBB1_3 +; RV32IBP-NEXT: .LBB1_2: +; RV32IBP-NEXT: not a1, a1 +; RV32IBP-NEXT: sll a1, a1, a2 +; RV32IBP-NEXT: addi a3, zero, 31 +; RV32IBP-NEXT: sub a3, a3, a2 +; RV32IBP-NEXT: srli a4, a0, 1 +; RV32IBP-NEXT: srl a3, a4, a3 +; RV32IBP-NEXT: or a1, a1, a3 +; RV32IBP-NEXT: sll a2, a0, a2 +; RV32IBP-NEXT: .LBB1_3: +; RV32IBP-NEXT: not a1, a1 +; RV32IBP-NEXT: not a0, a2 +; RV32IBP-NEXT: ret + %neg = xor i64 %a, -1 + %shl = shl i64 %neg, %b + %neg1 = xor i64 %shl, -1 + ret i64 %neg1 +} + +define i32 @sro_i32(i32 %a, i32 %b) nounwind { +; RV32I-LABEL: sro_i32: +; RV32I: # %bb.0: +; RV32I-NEXT: not a0, a0 +; RV32I-NEXT: srl a0, a0, a1 +; RV32I-NEXT: not a0, a0 +; RV32I-NEXT: ret +; +; RV32IB-LABEL: sro_i32: +; RV32IB: # %bb.0: +; RV32IB-NEXT: sro a0, a0, a1 +; RV32IB-NEXT: ret +; +; RV32IBP-LABEL: sro_i32: +; RV32IBP: # %bb.0: +; RV32IBP-NEXT: sro a0, a0, a1 +; RV32IBP-NEXT: ret + %neg = xor i32 %a, -1 + %shr = lshr i32 %neg, %b + %neg1 = xor i32 %shr, -1 + ret i32 %neg1 +} + +; As we are not matching directly i64 code patterns on RV32 some i64 patterns +; don't have yet any matching bit manipulation instructions on RV32. +; This test is presented here in case future expansions of the experimental-b +; extension introduce instructions suitable for this pattern. + +define i64 @sro_i64(i64 %a, i64 %b) nounwind { +; RV32I-LABEL: sro_i64: +; RV32I: # %bb.0: +; RV32I-NEXT: addi a3, a2, -32 +; RV32I-NEXT: not a1, a1 +; RV32I-NEXT: bltz a3, .LBB3_2 +; RV32I-NEXT: # %bb.1: +; RV32I-NEXT: mv a2, zero +; RV32I-NEXT: srl a0, a1, a3 +; RV32I-NEXT: j .LBB3_3 +; RV32I-NEXT: .LBB3_2: +; RV32I-NEXT: not a0, a0 +; RV32I-NEXT: srl a0, a0, a2 +; RV32I-NEXT: addi a3, zero, 31 +; RV32I-NEXT: sub a3, a3, a2 +; RV32I-NEXT: slli a4, a1, 1 +; RV32I-NEXT: sll a3, a4, a3 +; RV32I-NEXT: or a0, a0, a3 +; RV32I-NEXT: srl a2, a1, a2 +; RV32I-NEXT: .LBB3_3: +; RV32I-NEXT: not a0, a0 +; RV32I-NEXT: not a1, a2 +; RV32I-NEXT: ret +; +; RV32IB-LABEL: sro_i64: +; RV32IB: # %bb.0: +; RV32IB-NEXT: not a1, a1 +; RV32IB-NEXT: not a0, a0 +; RV32IB-NEXT: srl a0, a0, a2 +; RV32IB-NEXT: addi a3, zero, 31 +; RV32IB-NEXT: sub a3, a3, a2 +; RV32IB-NEXT: slli a4, a1, 1 +; RV32IB-NEXT: sll a3, a4, a3 +; RV32IB-NEXT: or a0, a0, a3 +; RV32IB-NEXT: addi a3, a2, -32 +; RV32IB-NEXT: srl a4, a1, a3 +; RV32IB-NEXT: slti a5, a3, 0 +; RV32IB-NEXT: cmov a0, a5, a0, a4 +; RV32IB-NEXT: srl a1, a1, a2 +; RV32IB-NEXT: srai a2, a3, 31 +; RV32IB-NEXT: and a1, a2, a1 +; RV32IB-NEXT: not a0, a0 +; RV32IB-NEXT: not a1, a1 +; RV32IB-NEXT: ret +; +; RV32IBP-LABEL: sro_i64: +; RV32IBP: # %bb.0: +; RV32IBP-NEXT: addi a3, a2, -32 +; RV32IBP-NEXT: not a1, a1 +; RV32IBP-NEXT: bltz a3, .LBB3_2 +; RV32IBP-NEXT: # %bb.1: +; RV32IBP-NEXT: mv a2, zero +; RV32IBP-NEXT: srl a0, a1, a3 +; RV32IBP-NEXT: j .LBB3_3 +; RV32IBP-NEXT: .LBB3_2: +; RV32IBP-NEXT: not a0, a0 +; RV32IBP-NEXT: srl a0, a0, a2 +; RV32IBP-NEXT: addi a3, zero, 31 +; RV32IBP-NEXT: sub a3, a3, a2 +; RV32IBP-NEXT: slli a4, a1, 1 +; RV32IBP-NEXT: sll a3, a4, a3 +; RV32IBP-NEXT: or a0, a0, a3 +; RV32IBP-NEXT: srl a2, a1, a2 +; RV32IBP-NEXT: .LBB3_3: +; RV32IBP-NEXT: not a0, a0 +; RV32IBP-NEXT: not a1, a2 +; RV32IBP-NEXT: ret + %neg = xor i64 %a, -1 + %shr = lshr i64 %neg, %b + %neg1 = xor i64 %shr, -1 + ret i64 %neg1 +} + +define i32 @sloi_i32(i32 %a) nounwind { +; RV32I-LABEL: sloi_i32: +; RV32I: # %bb.0: +; RV32I-NEXT: slli a0, a0, 1 +; RV32I-NEXT: ori a0, a0, 1 +; RV32I-NEXT: ret +; +; RV32IB-LABEL: sloi_i32: +; RV32IB: # %bb.0: +; RV32IB-NEXT: sloi a0, a0, 1 +; RV32IB-NEXT: ret +; +; RV32IBP-LABEL: sloi_i32: +; RV32IBP: # %bb.0: +; RV32IBP-NEXT: sloi a0, a0, 1 +; RV32IBP-NEXT: ret + %neg = shl i32 %a, 1 + %neg12 = or i32 %neg, 1 + ret i32 %neg12 +} + +define i64 @sloi_i64(i64 %a) nounwind { +; RV32I-LABEL: sloi_i64: +; RV32I: # %bb.0: +; RV32I-NEXT: srli a2, a0, 31 +; RV32I-NEXT: slli a1, a1, 1 +; RV32I-NEXT: or a1, a1, a2 +; RV32I-NEXT: slli a0, a0, 1 +; RV32I-NEXT: ori a0, a0, 1 +; RV32I-NEXT: ret +; +; RV32IB-LABEL: sloi_i64: +; RV32IB: # %bb.0: +; RV32IB-NEXT: fsri a1, a0, a1, 31 +; RV32IB-NEXT: sloi a0, a0, 1 +; RV32IB-NEXT: ret +; +; RV32IBP-LABEL: sloi_i64: +; RV32IBP: # %bb.0: +; RV32IBP-NEXT: srli a2, a0, 31 +; RV32IBP-NEXT: slli a1, a1, 1 +; RV32IBP-NEXT: or a1, a1, a2 +; RV32IBP-NEXT: sloi a0, a0, 1 +; RV32IBP-NEXT: ret + %neg = shl i64 %a, 1 + %neg12 = or i64 %neg, 1 + ret i64 %neg12 +} + +define i32 @sroi_i32(i32 %a) nounwind { +; RV32I-LABEL: sroi_i32: +; RV32I: # %bb.0: +; RV32I-NEXT: srli a0, a0, 1 +; RV32I-NEXT: lui a1, 524288 +; RV32I-NEXT: or a0, a0, a1 +; RV32I-NEXT: ret +; +; RV32IB-LABEL: sroi_i32: +; RV32IB: # %bb.0: +; RV32IB-NEXT: sroi a0, a0, 1 +; RV32IB-NEXT: ret +; +; RV32IBP-LABEL: sroi_i32: +; RV32IBP: # %bb.0: +; RV32IBP-NEXT: sroi a0, a0, 1 +; RV32IBP-NEXT: ret + %neg = lshr i32 %a, 1 + %neg12 = or i32 %neg, -2147483648 + ret i32 %neg12 +} + +define i64 @sroi_i64(i64 %a) nounwind { +; RV32I-LABEL: sroi_i64: +; RV32I: # %bb.0: +; RV32I-NEXT: slli a2, a1, 31 +; RV32I-NEXT: srli a0, a0, 1 +; RV32I-NEXT: or a0, a0, a2 +; RV32I-NEXT: srli a1, a1, 1 +; RV32I-NEXT: lui a2, 524288 +; RV32I-NEXT: or a1, a1, a2 +; RV32I-NEXT: ret +; +; RV32IB-LABEL: sroi_i64: +; RV32IB: # %bb.0: +; RV32IB-NEXT: fsri a0, a0, a1, 1 +; RV32IB-NEXT: sroi a1, a1, 1 +; RV32IB-NEXT: ret +; +; RV32IBP-LABEL: sroi_i64: +; RV32IBP: # %bb.0: +; RV32IBP-NEXT: slli a2, a1, 31 +; RV32IBP-NEXT: srli a0, a0, 1 +; RV32IBP-NEXT: or a0, a0, a2 +; RV32IBP-NEXT: sroi a1, a1, 1 +; RV32IBP-NEXT: ret + %neg = lshr i64 %a, 1 + %neg12 = or i64 %neg, -9223372036854775808 + ret i64 %neg12 +} + define i32 @gorc1_i32(i32 %a) nounwind { ; RV32I-LABEL: gorc1_i32: ; RV32I: # %bb.0: diff --git a/llvm/test/CodeGen/RISCV/rv64Zbb.ll b/llvm/test/CodeGen/RISCV/rv64Zbb.ll --- a/llvm/test/CodeGen/RISCV/rv64Zbb.ll +++ b/llvm/test/CodeGen/RISCV/rv64Zbb.ll @@ -6,216 +6,6 @@ ; RUN: llc -mtriple=riscv64 -mattr=+experimental-zbb -verify-machineinstrs < %s \ ; RUN: | FileCheck %s -check-prefix=RV64IBB -define signext i32 @slo_i32(i32 signext %a, i32 signext %b) nounwind { -; RV64I-LABEL: slo_i32: -; RV64I: # %bb.0: -; RV64I-NEXT: not a0, a0 -; RV64I-NEXT: sllw a0, a0, a1 -; RV64I-NEXT: not a0, a0 -; RV64I-NEXT: ret -; -; RV64IB-LABEL: slo_i32: -; RV64IB: # %bb.0: -; RV64IB-NEXT: slow a0, a0, a1 -; RV64IB-NEXT: ret -; -; RV64IBB-LABEL: slo_i32: -; RV64IBB: # %bb.0: -; RV64IBB-NEXT: slow a0, a0, a1 -; RV64IBB-NEXT: ret - %neg = xor i32 %a, -1 - %shl = shl i32 %neg, %b - %neg1 = xor i32 %shl, -1 - ret i32 %neg1 -} - -define i64 @slo_i64(i64 %a, i64 %b) nounwind { -; RV64I-LABEL: slo_i64: -; RV64I: # %bb.0: -; RV64I-NEXT: not a0, a0 -; RV64I-NEXT: sll a0, a0, a1 -; RV64I-NEXT: not a0, a0 -; RV64I-NEXT: ret -; -; RV64IB-LABEL: slo_i64: -; RV64IB: # %bb.0: -; RV64IB-NEXT: slo a0, a0, a1 -; RV64IB-NEXT: ret -; -; RV64IBB-LABEL: slo_i64: -; RV64IBB: # %bb.0: -; RV64IBB-NEXT: slo a0, a0, a1 -; RV64IBB-NEXT: ret - %neg = xor i64 %a, -1 - %shl = shl i64 %neg, %b - %neg1 = xor i64 %shl, -1 - ret i64 %neg1 -} - -define signext i32 @sro_i32(i32 signext %a, i32 signext %b) nounwind { -; RV64I-LABEL: sro_i32: -; RV64I: # %bb.0: -; RV64I-NEXT: not a0, a0 -; RV64I-NEXT: srlw a0, a0, a1 -; RV64I-NEXT: not a0, a0 -; RV64I-NEXT: ret -; -; RV64IB-LABEL: sro_i32: -; RV64IB: # %bb.0: -; RV64IB-NEXT: srow a0, a0, a1 -; RV64IB-NEXT: ret -; -; RV64IBB-LABEL: sro_i32: -; RV64IBB: # %bb.0: -; RV64IBB-NEXT: srow a0, a0, a1 -; RV64IBB-NEXT: ret - %neg = xor i32 %a, -1 - %shr = lshr i32 %neg, %b - %neg1 = xor i32 %shr, -1 - ret i32 %neg1 -} - -define i64 @sro_i64(i64 %a, i64 %b) nounwind { -; RV64I-LABEL: sro_i64: -; RV64I: # %bb.0: -; RV64I-NEXT: not a0, a0 -; RV64I-NEXT: srl a0, a0, a1 -; RV64I-NEXT: not a0, a0 -; RV64I-NEXT: ret -; -; RV64IB-LABEL: sro_i64: -; RV64IB: # %bb.0: -; RV64IB-NEXT: sro a0, a0, a1 -; RV64IB-NEXT: ret -; -; RV64IBB-LABEL: sro_i64: -; RV64IBB: # %bb.0: -; RV64IBB-NEXT: sro a0, a0, a1 -; RV64IBB-NEXT: ret - %neg = xor i64 %a, -1 - %shr = lshr i64 %neg, %b - %neg1 = xor i64 %shr, -1 - ret i64 %neg1 -} - -define signext i32 @sloi_i32(i32 signext %a) nounwind { -; RV64I-LABEL: sloi_i32: -; RV64I: # %bb.0: -; RV64I-NEXT: slli a0, a0, 1 -; RV64I-NEXT: ori a0, a0, 1 -; RV64I-NEXT: sext.w a0, a0 -; RV64I-NEXT: ret -; -; RV64IB-LABEL: sloi_i32: -; RV64IB: # %bb.0: -; RV64IB-NEXT: sloiw a0, a0, 1 -; RV64IB-NEXT: ret -; -; RV64IBB-LABEL: sloi_i32: -; RV64IBB: # %bb.0: -; RV64IBB-NEXT: sloiw a0, a0, 1 -; RV64IBB-NEXT: ret - %neg = shl i32 %a, 1 - %neg12 = or i32 %neg, 1 - ret i32 %neg12 -} - -define i64 @sloi_i64(i64 %a) nounwind { -; RV64I-LABEL: sloi_i64: -; RV64I: # %bb.0: -; RV64I-NEXT: slli a0, a0, 1 -; RV64I-NEXT: ori a0, a0, 1 -; RV64I-NEXT: ret -; -; RV64IB-LABEL: sloi_i64: -; RV64IB: # %bb.0: -; RV64IB-NEXT: sloi a0, a0, 1 -; RV64IB-NEXT: ret -; -; RV64IBB-LABEL: sloi_i64: -; RV64IBB: # %bb.0: -; RV64IBB-NEXT: sloi a0, a0, 1 -; RV64IBB-NEXT: ret - %neg = shl i64 %a, 1 - %neg12 = or i64 %neg, 1 - ret i64 %neg12 -} - -define signext i32 @sroi_i32(i32 signext %a) nounwind { -; RV64I-LABEL: sroi_i32: -; RV64I: # %bb.0: -; RV64I-NEXT: srli a0, a0, 1 -; RV64I-NEXT: lui a1, 524288 -; RV64I-NEXT: or a0, a0, a1 -; RV64I-NEXT: ret -; -; RV64IB-LABEL: sroi_i32: -; RV64IB: # %bb.0: -; RV64IB-NEXT: sroiw a0, a0, 1 -; RV64IB-NEXT: ret -; -; RV64IBB-LABEL: sroi_i32: -; RV64IBB: # %bb.0: -; RV64IBB-NEXT: sroiw a0, a0, 1 -; RV64IBB-NEXT: ret - %neg = lshr i32 %a, 1 - %neg12 = or i32 %neg, -2147483648 - ret i32 %neg12 -} - -; This is similar to the type legalized version of sroiw but the mask is 0 in -; the upper bits instead of 1 so the result is not sign extended. Make sure we -; don't match it to sroiw. -define i64 @sroiw_bug(i64 %a) nounwind { -; RV64I-LABEL: sroiw_bug: -; RV64I: # %bb.0: -; RV64I-NEXT: srli a0, a0, 1 -; RV64I-NEXT: addi a1, zero, 1 -; RV64I-NEXT: slli a1, a1, 31 -; RV64I-NEXT: or a0, a0, a1 -; RV64I-NEXT: ret -; -; RV64IB-LABEL: sroiw_bug: -; RV64IB: # %bb.0: -; RV64IB-NEXT: srli a0, a0, 1 -; RV64IB-NEXT: sbseti a0, a0, 31 -; RV64IB-NEXT: ret -; -; RV64IBB-LABEL: sroiw_bug: -; RV64IBB: # %bb.0: -; RV64IBB-NEXT: srli a0, a0, 1 -; RV64IBB-NEXT: addi a1, zero, 1 -; RV64IBB-NEXT: slli a1, a1, 31 -; RV64IBB-NEXT: or a0, a0, a1 -; RV64IBB-NEXT: ret - %neg = lshr i64 %a, 1 - %neg12 = or i64 %neg, 2147483648 - ret i64 %neg12 -} - -define i64 @sroi_i64(i64 %a) nounwind { -; RV64I-LABEL: sroi_i64: -; RV64I: # %bb.0: -; RV64I-NEXT: srli a0, a0, 1 -; RV64I-NEXT: addi a1, zero, -1 -; RV64I-NEXT: slli a1, a1, 63 -; RV64I-NEXT: or a0, a0, a1 -; RV64I-NEXT: ret -; -; RV64IB-LABEL: sroi_i64: -; RV64IB: # %bb.0: -; RV64IB-NEXT: sroi a0, a0, 1 -; RV64IB-NEXT: ret -; -; RV64IBB-LABEL: sroi_i64: -; RV64IBB: # %bb.0: -; RV64IBB-NEXT: sroi a0, a0, 1 -; RV64IBB-NEXT: ret - %neg = lshr i64 %a, 1 - %neg12 = or i64 %neg, -9223372036854775808 - ret i64 %neg12 -} - declare i32 @llvm.ctlz.i32(i32, i1) define signext i32 @ctlz_i32(i32 signext %a) nounwind { @@ -223,7 +13,7 @@ ; RV64I: # %bb.0: ; RV64I-NEXT: addi sp, sp, -16 ; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill -; RV64I-NEXT: beqz a0, .LBB9_2 +; RV64I-NEXT: beqz a0, .LBB0_2 ; RV64I-NEXT: # %bb.1: # %cond.false ; RV64I-NEXT: srliw a1, a0, 1 ; RV64I-NEXT: slli a0, a0, 32 @@ -283,10 +73,10 @@ ; RV64I-NEXT: call __muldi3@plt ; RV64I-NEXT: srli a0, a0, 56 ; RV64I-NEXT: addi a0, a0, -32 -; RV64I-NEXT: j .LBB9_3 -; RV64I-NEXT: .LBB9_2: +; RV64I-NEXT: j .LBB0_3 +; RV64I-NEXT: .LBB0_2: ; RV64I-NEXT: addi a0, zero, 32 -; RV64I-NEXT: .LBB9_3: # %cond.end +; RV64I-NEXT: .LBB0_3: # %cond.end ; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload ; RV64I-NEXT: addi sp, sp, 16 ; RV64I-NEXT: ret @@ -311,7 +101,7 @@ ; RV64I: # %bb.0: ; RV64I-NEXT: addi sp, sp, -16 ; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill -; RV64I-NEXT: beqz a0, .LBB10_2 +; RV64I-NEXT: beqz a0, .LBB1_2 ; RV64I-NEXT: # %bb.1: # %cond.false ; RV64I-NEXT: srli a1, a0, 1 ; RV64I-NEXT: or a0, a0, a1 @@ -368,10 +158,10 @@ ; RV64I-NEXT: addi a1, a1, 257 ; RV64I-NEXT: call __muldi3@plt ; RV64I-NEXT: srli a0, a0, 56 -; RV64I-NEXT: j .LBB10_3 -; RV64I-NEXT: .LBB10_2: +; RV64I-NEXT: j .LBB1_3 +; RV64I-NEXT: .LBB1_2: ; RV64I-NEXT: addi a0, zero, 64 -; RV64I-NEXT: .LBB10_3: # %cond.end +; RV64I-NEXT: .LBB1_3: # %cond.end ; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload ; RV64I-NEXT: addi sp, sp, 16 ; RV64I-NEXT: ret @@ -396,7 +186,7 @@ ; RV64I: # %bb.0: ; RV64I-NEXT: addi sp, sp, -16 ; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill -; RV64I-NEXT: beqz a0, .LBB11_2 +; RV64I-NEXT: beqz a0, .LBB2_2 ; RV64I-NEXT: # %bb.1: # %cond.false ; RV64I-NEXT: addi a1, a0, -1 ; RV64I-NEXT: not a0, a0 @@ -443,10 +233,10 @@ ; RV64I-NEXT: addi a1, a1, 257 ; RV64I-NEXT: call __muldi3@plt ; RV64I-NEXT: srli a0, a0, 56 -; RV64I-NEXT: j .LBB11_3 -; RV64I-NEXT: .LBB11_2: +; RV64I-NEXT: j .LBB2_3 +; RV64I-NEXT: .LBB2_2: ; RV64I-NEXT: addi a0, zero, 32 -; RV64I-NEXT: .LBB11_3: # %cond.end +; RV64I-NEXT: .LBB2_3: # %cond.end ; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload ; RV64I-NEXT: addi sp, sp, 16 ; RV64I-NEXT: ret @@ -471,7 +261,7 @@ ; RV64I: # %bb.0: ; RV64I-NEXT: addi sp, sp, -16 ; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill -; RV64I-NEXT: beqz a0, .LBB12_2 +; RV64I-NEXT: beqz a0, .LBB3_2 ; RV64I-NEXT: # %bb.1: # %cond.false ; RV64I-NEXT: addi a1, a0, -1 ; RV64I-NEXT: not a0, a0 @@ -518,10 +308,10 @@ ; RV64I-NEXT: addi a1, a1, 257 ; RV64I-NEXT: call __muldi3@plt ; RV64I-NEXT: srli a0, a0, 56 -; RV64I-NEXT: j .LBB12_3 -; RV64I-NEXT: .LBB12_2: +; RV64I-NEXT: j .LBB3_3 +; RV64I-NEXT: .LBB3_2: ; RV64I-NEXT: addi a0, zero, 64 -; RV64I-NEXT: .LBB12_3: # %cond.end +; RV64I-NEXT: .LBB3_3: # %cond.end ; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload ; RV64I-NEXT: addi sp, sp, 16 ; RV64I-NEXT: ret @@ -754,10 +544,10 @@ define signext i32 @min_i32(i32 signext %a, i32 signext %b) nounwind { ; RV64I-LABEL: min_i32: ; RV64I: # %bb.0: -; RV64I-NEXT: blt a0, a1, .LBB19_2 +; RV64I-NEXT: blt a0, a1, .LBB10_2 ; RV64I-NEXT: # %bb.1: ; RV64I-NEXT: mv a0, a1 -; RV64I-NEXT: .LBB19_2: +; RV64I-NEXT: .LBB10_2: ; RV64I-NEXT: ret ; ; RV64IB-LABEL: min_i32: @@ -777,10 +567,10 @@ define i64 @min_i64(i64 %a, i64 %b) nounwind { ; RV64I-LABEL: min_i64: ; RV64I: # %bb.0: -; RV64I-NEXT: blt a0, a1, .LBB20_2 +; RV64I-NEXT: blt a0, a1, .LBB11_2 ; RV64I-NEXT: # %bb.1: ; RV64I-NEXT: mv a0, a1 -; RV64I-NEXT: .LBB20_2: +; RV64I-NEXT: .LBB11_2: ; RV64I-NEXT: ret ; ; RV64IB-LABEL: min_i64: @@ -800,10 +590,10 @@ define signext i32 @max_i32(i32 signext %a, i32 signext %b) nounwind { ; RV64I-LABEL: max_i32: ; RV64I: # %bb.0: -; RV64I-NEXT: blt a1, a0, .LBB21_2 +; RV64I-NEXT: blt a1, a0, .LBB12_2 ; RV64I-NEXT: # %bb.1: ; RV64I-NEXT: mv a0, a1 -; RV64I-NEXT: .LBB21_2: +; RV64I-NEXT: .LBB12_2: ; RV64I-NEXT: ret ; ; RV64IB-LABEL: max_i32: @@ -823,10 +613,10 @@ define i64 @max_i64(i64 %a, i64 %b) nounwind { ; RV64I-LABEL: max_i64: ; RV64I: # %bb.0: -; RV64I-NEXT: blt a1, a0, .LBB22_2 +; RV64I-NEXT: blt a1, a0, .LBB13_2 ; RV64I-NEXT: # %bb.1: ; RV64I-NEXT: mv a0, a1 -; RV64I-NEXT: .LBB22_2: +; RV64I-NEXT: .LBB13_2: ; RV64I-NEXT: ret ; ; RV64IB-LABEL: max_i64: @@ -846,10 +636,10 @@ define signext i32 @minu_i32(i32 signext %a, i32 signext %b) nounwind { ; RV64I-LABEL: minu_i32: ; RV64I: # %bb.0: -; RV64I-NEXT: bltu a0, a1, .LBB23_2 +; RV64I-NEXT: bltu a0, a1, .LBB14_2 ; RV64I-NEXT: # %bb.1: ; RV64I-NEXT: mv a0, a1 -; RV64I-NEXT: .LBB23_2: +; RV64I-NEXT: .LBB14_2: ; RV64I-NEXT: ret ; ; RV64IB-LABEL: minu_i32: @@ -869,10 +659,10 @@ define i64 @minu_i64(i64 %a, i64 %b) nounwind { ; RV64I-LABEL: minu_i64: ; RV64I: # %bb.0: -; RV64I-NEXT: bltu a0, a1, .LBB24_2 +; RV64I-NEXT: bltu a0, a1, .LBB15_2 ; RV64I-NEXT: # %bb.1: ; RV64I-NEXT: mv a0, a1 -; RV64I-NEXT: .LBB24_2: +; RV64I-NEXT: .LBB15_2: ; RV64I-NEXT: ret ; ; RV64IB-LABEL: minu_i64: @@ -892,10 +682,10 @@ define signext i32 @maxu_i32(i32 signext %a, i32 signext %b) nounwind { ; RV64I-LABEL: maxu_i32: ; RV64I: # %bb.0: -; RV64I-NEXT: bltu a1, a0, .LBB25_2 +; RV64I-NEXT: bltu a1, a0, .LBB16_2 ; RV64I-NEXT: # %bb.1: ; RV64I-NEXT: mv a0, a1 -; RV64I-NEXT: .LBB25_2: +; RV64I-NEXT: .LBB16_2: ; RV64I-NEXT: ret ; ; RV64IB-LABEL: maxu_i32: @@ -915,10 +705,10 @@ define i64 @maxu_i64(i64 %a, i64 %b) nounwind { ; RV64I-LABEL: maxu_i64: ; RV64I: # %bb.0: -; RV64I-NEXT: bltu a1, a0, .LBB26_2 +; RV64I-NEXT: bltu a1, a0, .LBB17_2 ; RV64I-NEXT: # %bb.1: ; RV64I-NEXT: mv a0, a1 -; RV64I-NEXT: .LBB26_2: +; RV64I-NEXT: .LBB17_2: ; RV64I-NEXT: ret ; ; RV64IB-LABEL: maxu_i64: diff --git a/llvm/test/CodeGen/RISCV/rv64Zbp.ll b/llvm/test/CodeGen/RISCV/rv64Zbp.ll --- a/llvm/test/CodeGen/RISCV/rv64Zbp.ll +++ b/llvm/test/CodeGen/RISCV/rv64Zbp.ll @@ -6,6 +6,216 @@ ; RUN: llc -mtriple=riscv64 -mattr=+experimental-zbp -verify-machineinstrs < %s \ ; RUN: | FileCheck %s -check-prefix=RV64IBP +define signext i32 @slo_i32(i32 signext %a, i32 signext %b) nounwind { +; RV64I-LABEL: slo_i32: +; RV64I: # %bb.0: +; RV64I-NEXT: not a0, a0 +; RV64I-NEXT: sllw a0, a0, a1 +; RV64I-NEXT: not a0, a0 +; RV64I-NEXT: ret +; +; RV64IB-LABEL: slo_i32: +; RV64IB: # %bb.0: +; RV64IB-NEXT: slow a0, a0, a1 +; RV64IB-NEXT: ret +; +; RV64IBP-LABEL: slo_i32: +; RV64IBP: # %bb.0: +; RV64IBP-NEXT: slow a0, a0, a1 +; RV64IBP-NEXT: ret + %neg = xor i32 %a, -1 + %shl = shl i32 %neg, %b + %neg1 = xor i32 %shl, -1 + ret i32 %neg1 +} + +define i64 @slo_i64(i64 %a, i64 %b) nounwind { +; RV64I-LABEL: slo_i64: +; RV64I: # %bb.0: +; RV64I-NEXT: not a0, a0 +; RV64I-NEXT: sll a0, a0, a1 +; RV64I-NEXT: not a0, a0 +; RV64I-NEXT: ret +; +; RV64IB-LABEL: slo_i64: +; RV64IB: # %bb.0: +; RV64IB-NEXT: slo a0, a0, a1 +; RV64IB-NEXT: ret +; +; RV64IBP-LABEL: slo_i64: +; RV64IBP: # %bb.0: +; RV64IBP-NEXT: slo a0, a0, a1 +; RV64IBP-NEXT: ret + %neg = xor i64 %a, -1 + %shl = shl i64 %neg, %b + %neg1 = xor i64 %shl, -1 + ret i64 %neg1 +} + +define signext i32 @sro_i32(i32 signext %a, i32 signext %b) nounwind { +; RV64I-LABEL: sro_i32: +; RV64I: # %bb.0: +; RV64I-NEXT: not a0, a0 +; RV64I-NEXT: srlw a0, a0, a1 +; RV64I-NEXT: not a0, a0 +; RV64I-NEXT: ret +; +; RV64IB-LABEL: sro_i32: +; RV64IB: # %bb.0: +; RV64IB-NEXT: srow a0, a0, a1 +; RV64IB-NEXT: ret +; +; RV64IBP-LABEL: sro_i32: +; RV64IBP: # %bb.0: +; RV64IBP-NEXT: srow a0, a0, a1 +; RV64IBP-NEXT: ret + %neg = xor i32 %a, -1 + %shr = lshr i32 %neg, %b + %neg1 = xor i32 %shr, -1 + ret i32 %neg1 +} + +define i64 @sro_i64(i64 %a, i64 %b) nounwind { +; RV64I-LABEL: sro_i64: +; RV64I: # %bb.0: +; RV64I-NEXT: not a0, a0 +; RV64I-NEXT: srl a0, a0, a1 +; RV64I-NEXT: not a0, a0 +; RV64I-NEXT: ret +; +; RV64IB-LABEL: sro_i64: +; RV64IB: # %bb.0: +; RV64IB-NEXT: sro a0, a0, a1 +; RV64IB-NEXT: ret +; +; RV64IBP-LABEL: sro_i64: +; RV64IBP: # %bb.0: +; RV64IBP-NEXT: sro a0, a0, a1 +; RV64IBP-NEXT: ret + %neg = xor i64 %a, -1 + %shr = lshr i64 %neg, %b + %neg1 = xor i64 %shr, -1 + ret i64 %neg1 +} + +define signext i32 @sloi_i32(i32 signext %a) nounwind { +; RV64I-LABEL: sloi_i32: +; RV64I: # %bb.0: +; RV64I-NEXT: slli a0, a0, 1 +; RV64I-NEXT: ori a0, a0, 1 +; RV64I-NEXT: sext.w a0, a0 +; RV64I-NEXT: ret +; +; RV64IB-LABEL: sloi_i32: +; RV64IB: # %bb.0: +; RV64IB-NEXT: sloiw a0, a0, 1 +; RV64IB-NEXT: ret +; +; RV64IBP-LABEL: sloi_i32: +; RV64IBP: # %bb.0: +; RV64IBP-NEXT: sloiw a0, a0, 1 +; RV64IBP-NEXT: ret + %neg = shl i32 %a, 1 + %neg12 = or i32 %neg, 1 + ret i32 %neg12 +} + +define i64 @sloi_i64(i64 %a) nounwind { +; RV64I-LABEL: sloi_i64: +; RV64I: # %bb.0: +; RV64I-NEXT: slli a0, a0, 1 +; RV64I-NEXT: ori a0, a0, 1 +; RV64I-NEXT: ret +; +; RV64IB-LABEL: sloi_i64: +; RV64IB: # %bb.0: +; RV64IB-NEXT: sloi a0, a0, 1 +; RV64IB-NEXT: ret +; +; RV64IBP-LABEL: sloi_i64: +; RV64IBP: # %bb.0: +; RV64IBP-NEXT: sloi a0, a0, 1 +; RV64IBP-NEXT: ret + %neg = shl i64 %a, 1 + %neg12 = or i64 %neg, 1 + ret i64 %neg12 +} + +define signext i32 @sroi_i32(i32 signext %a) nounwind { +; RV64I-LABEL: sroi_i32: +; RV64I: # %bb.0: +; RV64I-NEXT: srli a0, a0, 1 +; RV64I-NEXT: lui a1, 524288 +; RV64I-NEXT: or a0, a0, a1 +; RV64I-NEXT: ret +; +; RV64IB-LABEL: sroi_i32: +; RV64IB: # %bb.0: +; RV64IB-NEXT: sroiw a0, a0, 1 +; RV64IB-NEXT: ret +; +; RV64IBP-LABEL: sroi_i32: +; RV64IBP: # %bb.0: +; RV64IBP-NEXT: sroiw a0, a0, 1 +; RV64IBP-NEXT: ret + %neg = lshr i32 %a, 1 + %neg12 = or i32 %neg, -2147483648 + ret i32 %neg12 +} + +; This is similar to the type legalized version of sroiw but the mask is 0 in +; the upper bits instead of 1 so the result is not sign extended. Make sure we +; don't match it to sroiw. +define i64 @sroiw_bug(i64 %a) nounwind { +; RV64I-LABEL: sroiw_bug: +; RV64I: # %bb.0: +; RV64I-NEXT: srli a0, a0, 1 +; RV64I-NEXT: addi a1, zero, 1 +; RV64I-NEXT: slli a1, a1, 31 +; RV64I-NEXT: or a0, a0, a1 +; RV64I-NEXT: ret +; +; RV64IB-LABEL: sroiw_bug: +; RV64IB: # %bb.0: +; RV64IB-NEXT: srli a0, a0, 1 +; RV64IB-NEXT: sbseti a0, a0, 31 +; RV64IB-NEXT: ret +; +; RV64IBP-LABEL: sroiw_bug: +; RV64IBP: # %bb.0: +; RV64IBP-NEXT: srli a0, a0, 1 +; RV64IBP-NEXT: addi a1, zero, 1 +; RV64IBP-NEXT: slli a1, a1, 31 +; RV64IBP-NEXT: or a0, a0, a1 +; RV64IBP-NEXT: ret + %neg = lshr i64 %a, 1 + %neg12 = or i64 %neg, 2147483648 + ret i64 %neg12 +} + +define i64 @sroi_i64(i64 %a) nounwind { +; RV64I-LABEL: sroi_i64: +; RV64I: # %bb.0: +; RV64I-NEXT: srli a0, a0, 1 +; RV64I-NEXT: addi a1, zero, -1 +; RV64I-NEXT: slli a1, a1, 63 +; RV64I-NEXT: or a0, a0, a1 +; RV64I-NEXT: ret +; +; RV64IB-LABEL: sroi_i64: +; RV64IB: # %bb.0: +; RV64IB-NEXT: sroi a0, a0, 1 +; RV64IB-NEXT: ret +; +; RV64IBP-LABEL: sroi_i64: +; RV64IBP: # %bb.0: +; RV64IBP-NEXT: sroi a0, a0, 1 +; RV64IBP-NEXT: ret + %neg = lshr i64 %a, 1 + %neg12 = or i64 %neg, -9223372036854775808 + ret i64 %neg12 +} + define signext i32 @gorc1_i32(i32 signext %a) nounwind { ; RV64I-LABEL: gorc1_i32: ; RV64I: # %bb.0: diff --git a/llvm/test/MC/RISCV/rv32zbb-invalid.s b/llvm/test/MC/RISCV/rv32zbb-invalid.s --- a/llvm/test/MC/RISCV/rv32zbb-invalid.s +++ b/llvm/test/MC/RISCV/rv32zbb-invalid.s @@ -1,19 +1,5 @@ # RUN: not llvm-mc -triple riscv32 -mattr=+experimental-b,experimental-zbb < %s 2>&1 | FileCheck %s -# Too few operands -slo t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction -# Too few operands -sro t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction -# Too few operands -sloi t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction -# Immediate operand out of range -sloi t0, t1, 32 # CHECK: :[[@LINE]]:14: error: immediate must be an integer in the range [0, 31] -sloi t0, t1, -1 # CHECK: :[[@LINE]]:14: error: immediate must be an integer in the range [0, 31] -# Too few operands -sroi t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction -# Immediate operand out of range -sroi t0, t1, 32 # CHECK: :[[@LINE]]:14: error: immediate must be an integer in the range [0, 31] -sroi t0, t1, -1 # CHECK: :[[@LINE]]:14: error: immediate must be an integer in the range [0, 31] # Too many operands clz t0, t1, t2 # CHECK: :[[@LINE]]:13: error: invalid operand for instruction # Too many operands diff --git a/llvm/test/MC/RISCV/rv32zbb-valid.s b/llvm/test/MC/RISCV/rv32zbb-valid.s --- a/llvm/test/MC/RISCV/rv32zbb-valid.s +++ b/llvm/test/MC/RISCV/rv32zbb-valid.s @@ -12,18 +12,6 @@ # RUN: | llvm-objdump --mattr=+experimental-zbb -d -r - \ # RUN: | FileCheck --check-prefix=CHECK-ASM-AND-OBJ %s -# CHECK-ASM-AND-OBJ: slo t0, t1, t2 -# CHECK-ASM: encoding: [0xb3,0x12,0x73,0x20] -slo t0, t1, t2 -# CHECK-ASM-AND-OBJ: sro t0, t1, t2 -# CHECK-ASM: encoding: [0xb3,0x52,0x73,0x20] -sro t0, t1, t2 -# CHECK-ASM-AND-OBJ: sloi t0, t1, 0 -# CHECK-ASM: encoding: [0x93,0x12,0x03,0x20] -sloi t0, t1, 0 -# CHECK-ASM-AND-OBJ: sroi t0, t1, 0 -# CHECK-ASM: encoding: [0x93,0x52,0x03,0x20] -sroi t0, t1, 0 # CHECK-ASM-AND-OBJ: clz t0, t1 # CHECK-ASM: encoding: [0x93,0x12,0x03,0x60] clz t0, t1 diff --git a/llvm/test/MC/RISCV/rv32zbp-invalid.s b/llvm/test/MC/RISCV/rv32zbp-invalid.s --- a/llvm/test/MC/RISCV/rv32zbp-invalid.s +++ b/llvm/test/MC/RISCV/rv32zbp-invalid.s @@ -1,5 +1,19 @@ # RUN: not llvm-mc -triple riscv32 -mattr=+experimental-b,experimental-zbp < %s 2>&1 | FileCheck %s +# Too few operands +slo t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction +# Too few operands +sro t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction +# Too few operands +sloi t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction +# Immediate operand out of range +sloi t0, t1, 32 # CHECK: :[[@LINE]]:14: error: immediate must be an integer in the range [0, 31] +sloi t0, t1, -1 # CHECK: :[[@LINE]]:14: error: immediate must be an integer in the range [0, 31] +# Too few operands +sroi t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction +# Immediate operand out of range +sroi t0, t1, 32 # CHECK: :[[@LINE]]:14: error: immediate must be an integer in the range [0, 31] +sroi t0, t1, -1 # CHECK: :[[@LINE]]:14: error: immediate must be an integer in the range [0, 31] # Too few operands gorc t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction # Too few operands diff --git a/llvm/test/MC/RISCV/rv32zbp-valid.s b/llvm/test/MC/RISCV/rv32zbp-valid.s --- a/llvm/test/MC/RISCV/rv32zbp-valid.s +++ b/llvm/test/MC/RISCV/rv32zbp-valid.s @@ -12,6 +12,18 @@ # RUN: | llvm-objdump --mattr=+experimental-zbp -d -r - \ # RUN: | FileCheck --check-prefix=CHECK-ASM-AND-OBJ %s +# CHECK-ASM-AND-OBJ: slo t0, t1, t2 +# CHECK-ASM: encoding: [0xb3,0x12,0x73,0x20] +slo t0, t1, t2 +# CHECK-ASM-AND-OBJ: sro t0, t1, t2 +# CHECK-ASM: encoding: [0xb3,0x52,0x73,0x20] +sro t0, t1, t2 +# CHECK-ASM-AND-OBJ: sloi t0, t1, 0 +# CHECK-ASM: encoding: [0x93,0x12,0x03,0x20] +sloi t0, t1, 0 +# CHECK-ASM-AND-OBJ: sroi t0, t1, 0 +# CHECK-ASM: encoding: [0x93,0x52,0x03,0x20] +sroi t0, t1, 0 # CHECK-ASM-AND-OBJ: gorc t0, t1, t2 # CHECK-ASM: encoding: [0xb3,0x52,0x73,0x28] gorc t0, t1, t2 diff --git a/llvm/test/MC/RISCV/rv64zbb-invalid.s b/llvm/test/MC/RISCV/rv64zbb-invalid.s --- a/llvm/test/MC/RISCV/rv64zbb-invalid.s +++ b/llvm/test/MC/RISCV/rv64zbb-invalid.s @@ -1,19 +1,5 @@ # RUN: not llvm-mc -triple riscv64 -mattr=+experimental-b,experimental-zbb < %s 2>&1 | FileCheck %s -# Too few operands -slow t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction -# Too few operands -srow t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction -# Too few operands -sloiw t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction -# Immediate operand out of range -sloiw t0, t1, 32 # CHECK: :[[@LINE]]:15: error: immediate must be an integer in the range [0, 31] -sloiw t0, t1, -1 # CHECK: :[[@LINE]]:15: error: immediate must be an integer in the range [0, 31] -# Too few operands -sroiw t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction -# Immediate operand out of range -sroiw t0, t1, 32 # CHECK: :[[@LINE]]:15: error: immediate must be an integer in the range [0, 31] -sroiw t0, t1, -1 # CHECK: :[[@LINE]]:15: error: immediate must be an integer in the range [0, 31] # Too many operands clzw t0, t1, t2 # CHECK: :[[@LINE]]:14: error: invalid operand for instruction # Too many operands diff --git a/llvm/test/MC/RISCV/rv64zbb-valid.s b/llvm/test/MC/RISCV/rv64zbb-valid.s --- a/llvm/test/MC/RISCV/rv64zbb-valid.s +++ b/llvm/test/MC/RISCV/rv64zbb-valid.s @@ -12,18 +12,6 @@ # RUN: | llvm-objdump --mattr=+experimental-zbb -d -r - \ # RUN: | FileCheck --check-prefix=CHECK-ASM-AND-OBJ %s -# CHECK-ASM-AND-OBJ: slow t0, t1, t2 -# CHECK-ASM: encoding: [0xbb,0x12,0x73,0x20] -slow t0, t1, t2 -# CHECK-ASM-AND-OBJ: srow t0, t1, t2 -# CHECK-ASM: encoding: [0xbb,0x52,0x73,0x20] -srow t0, t1, t2 -# CHECK-ASM-AND-OBJ: sloiw t0, t1, 0 -# CHECK-ASM: encoding: [0x9b,0x12,0x03,0x20] -sloiw t0, t1, 0 -# CHECK-ASM-AND-OBJ: sroiw t0, t1, 0 -# CHECK-ASM: encoding: [0x9b,0x52,0x03,0x20] -sroiw t0, t1, 0 # CHECK-ASM-AND-OBJ: clzw t0, t1 # CHECK-ASM: encoding: [0x9b,0x12,0x03,0x60] clzw t0, t1 diff --git a/llvm/test/MC/RISCV/rv64zbp-invalid.s b/llvm/test/MC/RISCV/rv64zbp-invalid.s --- a/llvm/test/MC/RISCV/rv64zbp-invalid.s +++ b/llvm/test/MC/RISCV/rv64zbp-invalid.s @@ -1,5 +1,19 @@ # RUN: not llvm-mc -triple riscv64 -mattr=+experimental-b,experimental-zbp < %s 2>&1 | FileCheck %s +# Too few operands +slow t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction +# Too few operands +srow t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction +# Too few operands +sloiw t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction +# Immediate operand out of range +sloiw t0, t1, 32 # CHECK: :[[@LINE]]:15: error: immediate must be an integer in the range [0, 31] +sloiw t0, t1, -1 # CHECK: :[[@LINE]]:15: error: immediate must be an integer in the range [0, 31] +# Too few operands +sroiw t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction +# Immediate operand out of range +sroiw t0, t1, 32 # CHECK: :[[@LINE]]:15: error: immediate must be an integer in the range [0, 31] +sroiw t0, t1, -1 # CHECK: :[[@LINE]]:15: error: immediate must be an integer in the range [0, 31] # Too few operands gorcw t0, t1 # CHECK: :[[@LINE]]:1: error: too few operands for instruction # Too few operands diff --git a/llvm/test/MC/RISCV/rv64zbp-valid.s b/llvm/test/MC/RISCV/rv64zbp-valid.s --- a/llvm/test/MC/RISCV/rv64zbp-valid.s +++ b/llvm/test/MC/RISCV/rv64zbp-valid.s @@ -12,6 +12,18 @@ # RUN: | llvm-objdump --mattr=+experimental-zbp -d -r - \ # RUN: | FileCheck --check-prefix=CHECK-ASM-AND-OBJ %s +# CHECK-ASM-AND-OBJ: slow t0, t1, t2 +# CHECK-ASM: encoding: [0xbb,0x12,0x73,0x20] +slow t0, t1, t2 +# CHECK-ASM-AND-OBJ: srow t0, t1, t2 +# CHECK-ASM: encoding: [0xbb,0x52,0x73,0x20] +srow t0, t1, t2 +# CHECK-ASM-AND-OBJ: sloiw t0, t1, 0 +# CHECK-ASM: encoding: [0x9b,0x12,0x03,0x20] +sloiw t0, t1, 0 +# CHECK-ASM-AND-OBJ: sroiw t0, t1, 0 +# CHECK-ASM: encoding: [0x9b,0x52,0x03,0x20] +sroiw t0, t1, 0 # CHECK-ASM-AND-OBJ: gorcw t0, t1, t2 # CHECK-ASM: encoding: [0xbb,0x52,0x73,0x28] gorcw t0, t1, t2