1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=riscv32 -mattr=+f -verify-machineinstrs < %s \
; RUN: -target-abi=ilp32f | FileCheck -check-prefixes=CHECKIF,RV32IF %s
; RUN: llc -mtriple=riscv64 -mattr=+f -verify-machineinstrs < %s \
; RUN: -target-abi=lp64f | FileCheck -check-prefixes=CHECKIF,RV64IF %s
define dso_local float @flw(ptr %a) nounwind {
; CHECKIF-LABEL: flw:
; CHECKIF: # %bb.0:
; CHECKIF-NEXT: flw ft0, 0(a0)
; CHECKIF-NEXT: flw ft1, 12(a0)
; CHECKIF-NEXT: fadd.s fa0, ft0, ft1
; CHECKIF-NEXT: ret
%1 = load float, ptr %a
%2 = getelementptr float, ptr %a, i32 3
%3 = load float, ptr %2
; Use both loaded values in an FP op to ensure an flw is used, even for the
; soft float ABI
%4 = fadd float %1, %3
ret float %4
}
define dso_local void @fsw(ptr %a, float %b, float %c) nounwind {
; Use %b and %c in an FP op to ensure floating point registers are used, even
; for the soft float ABI
; CHECKIF-LABEL: fsw:
; CHECKIF: # %bb.0:
; CHECKIF-NEXT: fadd.s ft0, fa0, fa1
; CHECKIF-NEXT: fsw ft0, 0(a0)
; CHECKIF-NEXT: fsw ft0, 32(a0)
; CHECKIF-NEXT: ret
%1 = fadd float %b, %c
store float %1, ptr %a
%2 = getelementptr float, ptr %a, i32 8
store float %1, ptr %2
ret void
}
; Check load and store to a global
@G = dso_local global float 0.0
define dso_local float @flw_fsw_global(float %a, float %b) nounwind {
; Use %a and %b in an FP op to ensure floating point registers are used, even
; for the soft float ABI
; CHECKIF-LABEL: flw_fsw_global:
; CHECKIF: # %bb.0:
; CHECKIF-NEXT: fadd.s fa0, fa0, fa1
; CHECKIF-NEXT: lui a0, %hi(G)
; CHECKIF-NEXT: flw ft0, %lo(G)(a0)
; CHECKIF-NEXT: addi a1, a0, %lo(G)
; CHECKIF-NEXT: fsw fa0, %lo(G)(a0)
; CHECKIF-NEXT: flw ft0, 36(a1)
; CHECKIF-NEXT: fsw fa0, 36(a1)
; CHECKIF-NEXT: ret
%1 = fadd float %a, %b
%2 = load volatile float, ptr @G
store float %1, ptr @G
%3 = getelementptr float, ptr @G, i32 9
%4 = load volatile float, ptr %3
store float %1, ptr %3
ret float %1
}
; Ensure that 1 is added to the high 20 bits if bit 11 of the low part is 1
define dso_local float @flw_fsw_constant(float %a) nounwind {
; RV32IF-LABEL: flw_fsw_constant:
; RV32IF: # %bb.0:
; RV32IF-NEXT: lui a0, 912092
; RV32IF-NEXT: flw ft0, -273(a0)
; RV32IF-NEXT: fadd.s fa0, fa0, ft0
; RV32IF-NEXT: fsw fa0, -273(a0)
; RV32IF-NEXT: ret
;
; RV64IF-LABEL: flw_fsw_constant:
; RV64IF: # %bb.0:
; RV64IF-NEXT: lui a0, 228023
; RV64IF-NEXT: slli a0, a0, 2
; RV64IF-NEXT: flw ft0, -273(a0)
; RV64IF-NEXT: fadd.s fa0, fa0, ft0
; RV64IF-NEXT: fsw fa0, -273(a0)
; RV64IF-NEXT: ret
%1 = inttoptr i32 3735928559 to ptr
%2 = load volatile float, ptr %1
%3 = fadd float %a, %2
store float %3, ptr %1
ret float %3
}
declare void @notdead(ptr)
define dso_local float @flw_stack(float %a) nounwind {
; RV32IF-LABEL: flw_stack:
; RV32IF: # %bb.0:
; RV32IF-NEXT: addi sp, sp, -16
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
; RV32IF-NEXT: fsw fs0, 8(sp) # 4-byte Folded Spill
; RV32IF-NEXT: fmv.s fs0, fa0
; RV32IF-NEXT: addi a0, sp, 4
; RV32IF-NEXT: call notdead@plt
; RV32IF-NEXT: flw ft0, 4(sp)
; RV32IF-NEXT: fadd.s fa0, ft0, fs0
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
; RV32IF-NEXT: flw fs0, 8(sp) # 4-byte Folded Reload
; RV32IF-NEXT: addi sp, sp, 16
; RV32IF-NEXT: ret
;
; RV64IF-LABEL: flw_stack:
; RV64IF: # %bb.0:
; RV64IF-NEXT: addi sp, sp, -16
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
; RV64IF-NEXT: fsw fs0, 4(sp) # 4-byte Folded Spill
; RV64IF-NEXT: fmv.s fs0, fa0
; RV64IF-NEXT: mv a0, sp
; RV64IF-NEXT: call notdead@plt
; RV64IF-NEXT: flw ft0, 0(sp)
; RV64IF-NEXT: fadd.s fa0, ft0, fs0
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
; RV64IF-NEXT: flw fs0, 4(sp) # 4-byte Folded Reload
; RV64IF-NEXT: addi sp, sp, 16
; RV64IF-NEXT: ret
%1 = alloca float, align 4
call void @notdead(ptr %1)
%2 = load float, ptr %1
%3 = fadd float %2, %a ; force load in to FPR32
ret float %3
}
define dso_local void @fsw_stack(float %a, float %b) nounwind {
; RV32IF-LABEL: fsw_stack:
; RV32IF: # %bb.0:
; RV32IF-NEXT: addi sp, sp, -16
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
; RV32IF-NEXT: fadd.s ft0, fa0, fa1
; RV32IF-NEXT: fsw ft0, 8(sp)
; RV32IF-NEXT: addi a0, sp, 8
; RV32IF-NEXT: call notdead@plt
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
; RV32IF-NEXT: addi sp, sp, 16
; RV32IF-NEXT: ret
;
; RV64IF-LABEL: fsw_stack:
; RV64IF: # %bb.0:
; RV64IF-NEXT: addi sp, sp, -16
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
; RV64IF-NEXT: fadd.s ft0, fa0, fa1
; RV64IF-NEXT: fsw ft0, 4(sp)
; RV64IF-NEXT: addi a0, sp, 4
; RV64IF-NEXT: call notdead@plt
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
; RV64IF-NEXT: addi sp, sp, 16
; RV64IF-NEXT: ret
%1 = fadd float %a, %b ; force store from FPR32
%2 = alloca float, align 4
store float %1, ptr %2
call void @notdead(ptr %2)
ret void
}
|