; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=riscv32 -mattr=+experimental-zfh -verify-machineinstrs \
; RUN:   -target-abi ilp32f < %s | FileCheck -check-prefix=RV32IZFH %s
; RUN: llc -mtriple=riscv64 -mattr=+experimental-zfh -verify-machineinstrs \
; RUN:   -target-abi lp64f < %s | FileCheck -check-prefix=RV64IZFH %s

define half @flh(half *%a) nounwind {
; RV32IZFH-LABEL: flh:
; RV32IZFH:       # %bb.0:
; RV32IZFH-NEXT:    flh ft0, 0(a0)
; RV32IZFH-NEXT:    flh ft1, 6(a0)
; RV32IZFH-NEXT:    fadd.h fa0, ft0, ft1
; RV32IZFH-NEXT:    ret
;
; RV64IZFH-LABEL: flh:
; RV64IZFH:       # %bb.0:
; RV64IZFH-NEXT:    flh ft0, 0(a0)
; RV64IZFH-NEXT:    flh ft1, 6(a0)
; RV64IZFH-NEXT:    fadd.h fa0, ft0, ft1
; RV64IZFH-NEXT:    ret
  %1 = load half, half* %a
  %2 = getelementptr half, half* %a, i32 3
  %3 = load half, half* %2
; Use both loaded values in an FP op to ensure an flh is used, even for the
; soft half ABI
  %4 = fadd half %1, %3
  ret half %4
}

define dso_local void @fsh(half *%a, half %b, half %c) nounwind {
; Use %b and %c in an FP op to ensure half precision floating point registers
; are used, even for the soft half ABI
; RV32IZFH-LABEL: fsh:
; RV32IZFH:       # %bb.0:
; RV32IZFH-NEXT:    fadd.h ft0, fa0, fa1
; RV32IZFH-NEXT:    fsh ft0, 0(a0)
; RV32IZFH-NEXT:    fsh ft0, 16(a0)
; RV32IZFH-NEXT:    ret
;
; RV64IZFH-LABEL: fsh:
; RV64IZFH:       # %bb.0:
; RV64IZFH-NEXT:    fadd.h ft0, fa0, fa1
; RV64IZFH-NEXT:    fsh ft0, 0(a0)
; RV64IZFH-NEXT:    fsh ft0, 16(a0)
; RV64IZFH-NEXT:    ret
  %1 = fadd half %b, %c
  store half %1, half* %a
  %2 = getelementptr half, half* %a, i32 8
  store half %1, half* %2
  ret void
}

; Check load and store to a global
@G = dso_local global half 0.0

define half @flh_fsh_global(half %a, half %b) nounwind {
; Use %a and %b in an FP op to ensure half precision floating point registers
; are used, even for the soft half ABI
; RV32IZFH-LABEL: flh_fsh_global:
; RV32IZFH:       # %bb.0:
; RV32IZFH-NEXT:    fadd.h fa0, fa0, fa1
; RV32IZFH-NEXT:    lui a0, %hi(G)
; RV32IZFH-NEXT:    flh ft0, %lo(G)(a0)
; RV32IZFH-NEXT:    fsh fa0, %lo(G)(a0)
; RV32IZFH-NEXT:    addi a0, a0, %lo(G)
; RV32IZFH-NEXT:    flh ft0, 18(a0)
; RV32IZFH-NEXT:    fsh fa0, 18(a0)
; RV32IZFH-NEXT:    ret
;
; RV64IZFH-LABEL: flh_fsh_global:
; RV64IZFH:       # %bb.0:
; RV64IZFH-NEXT:    fadd.h fa0, fa0, fa1
; RV64IZFH-NEXT:    lui a0, %hi(G)
; RV64IZFH-NEXT:    flh ft0, %lo(G)(a0)
; RV64IZFH-NEXT:    fsh fa0, %lo(G)(a0)
; RV64IZFH-NEXT:    addi a0, a0, %lo(G)
; RV64IZFH-NEXT:    flh ft0, 18(a0)
; RV64IZFH-NEXT:    fsh fa0, 18(a0)
; RV64IZFH-NEXT:    ret
  %1 = fadd half %a, %b
  %2 = load volatile half, half* @G
  store half %1, half* @G
  %3 = getelementptr half, half* @G, i32 9
  %4 = load volatile half, half* %3
  store half %1, half* %3
  ret half %1
}

; Ensure that 1 is added to the high 20 bits if bit 11 of the low part is 1
define half @flh_fsh_constant(half %a) nounwind {
; RV32IZFH-LABEL: flh_fsh_constant:
; RV32IZFH:       # %bb.0:
; RV32IZFH-NEXT:    lui a0, 912092
; RV32IZFH-NEXT:    flh ft0, -273(a0)
; RV32IZFH-NEXT:    fadd.h fa0, fa0, ft0
; RV32IZFH-NEXT:    fsh fa0, -273(a0)
; RV32IZFH-NEXT:    ret
;
; RV64IZFH-LABEL: flh_fsh_constant:
; RV64IZFH:       # %bb.0:
; RV64IZFH-NEXT:    lui a0, 228023
; RV64IZFH-NEXT:    slli a0, a0, 2
; RV64IZFH-NEXT:    flh ft0, -273(a0)
; RV64IZFH-NEXT:    fadd.h fa0, fa0, ft0
; RV64IZFH-NEXT:    fsh fa0, -273(a0)
; RV64IZFH-NEXT:    ret
  %1 = inttoptr i32 3735928559 to half*
  %2 = load volatile half, half* %1
  %3 = fadd half %a, %2
  store half %3, half* %1
  ret half %3
}

declare void @notdead(i8*)

define half @flh_stack(half %a) nounwind {
; RV32IZFH-LABEL: flh_stack:
; RV32IZFH:       # %bb.0:
; RV32IZFH-NEXT:    addi sp, sp, -16
; RV32IZFH-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
; RV32IZFH-NEXT:    fsw fs0, 8(sp) # 4-byte Folded Spill
; RV32IZFH-NEXT:    fmv.h fs0, fa0
; RV32IZFH-NEXT:    addi a0, sp, 4
; RV32IZFH-NEXT:    call notdead@plt
; RV32IZFH-NEXT:    flh ft0, 4(sp)
; RV32IZFH-NEXT:    fadd.h fa0, ft0, fs0
; RV32IZFH-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
; RV32IZFH-NEXT:    flw fs0, 8(sp) # 4-byte Folded Reload
; RV32IZFH-NEXT:    addi sp, sp, 16
; RV32IZFH-NEXT:    ret
;
; RV64IZFH-LABEL: flh_stack:
; RV64IZFH:       # %bb.0:
; RV64IZFH-NEXT:    addi sp, sp, -16
; RV64IZFH-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
; RV64IZFH-NEXT:    fsw fs0, 4(sp) # 4-byte Folded Spill
; RV64IZFH-NEXT:    fmv.h fs0, fa0
; RV64IZFH-NEXT:    mv a0, sp
; RV64IZFH-NEXT:    call notdead@plt
; RV64IZFH-NEXT:    flh ft0, 0(sp)
; RV64IZFH-NEXT:    fadd.h fa0, ft0, fs0
; RV64IZFH-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
; RV64IZFH-NEXT:    flw fs0, 4(sp) # 4-byte Folded Reload
; RV64IZFH-NEXT:    addi sp, sp, 16
; RV64IZFH-NEXT:    ret
  %1 = alloca half, align 4
  %2 = bitcast half* %1 to i8*
  call void @notdead(i8* %2)
  %3 = load half, half* %1
  %4 = fadd half %3, %a ; force load in to FPR16
  ret half %4
}

define dso_local void @fsh_stack(half %a, half %b) nounwind {
; RV32IZFH-LABEL: fsh_stack:
; RV32IZFH:       # %bb.0:
; RV32IZFH-NEXT:    addi sp, sp, -16
; RV32IZFH-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
; RV32IZFH-NEXT:    fadd.h ft0, fa0, fa1
; RV32IZFH-NEXT:    fsh ft0, 8(sp)
; RV32IZFH-NEXT:    addi a0, sp, 8
; RV32IZFH-NEXT:    call notdead@plt
; RV32IZFH-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
; RV32IZFH-NEXT:    addi sp, sp, 16
; RV32IZFH-NEXT:    ret
;
; RV64IZFH-LABEL: fsh_stack:
; RV64IZFH:       # %bb.0:
; RV64IZFH-NEXT:    addi sp, sp, -16
; RV64IZFH-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
; RV64IZFH-NEXT:    fadd.h ft0, fa0, fa1
; RV64IZFH-NEXT:    fsh ft0, 4(sp)
; RV64IZFH-NEXT:    addi a0, sp, 4
; RV64IZFH-NEXT:    call notdead@plt
; RV64IZFH-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
; RV64IZFH-NEXT:    addi sp, sp, 16
; RV64IZFH-NEXT:    ret
  %1 = fadd half %a, %b ; force store from FPR16
  %2 = alloca half, align 4
  store half %1, half* %2
  %3 = bitcast half* %2 to i8*
  call void @notdead(i8* %3)
  ret void
}