test/CodeGen/LoongArch/lsx/bitsel.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 3
; RUN: llc --mtriple=loongarch32 --mattr=+32s,+lsx < %s | FileCheck %s
; RUN: llc --mtriple=loongarch64 --mattr=+lsx < %s | FileCheck %s

define <16 x i8> @bitsel_v16i8(<16 x i8> %a, <16 x i8> %b, <16 x i8> %c) nounwind {
; CHECK-LABEL: bitsel_v16i8:
; CHECK:       # %bb.0: # %entry
; CHECK-NEXT:    vbitsel.v $vr0, $vr0, $vr1, $vr2
; CHECK-NEXT:    ret
entry:
  %0 = and <16 x i8> %c, %b
  %1 = xor <16 x i8> %c, splat (i8 -1)
  %2 = and <16 x i8> %a, %1
  %3 = or <16 x i8> %0, %2
  ret <16 x i8> %3
}

define <8 x i16> @bitsel_v8i16(<8 x i16> %a, <8 x i16> %b, <8 x i16> %c) nounwind {
; CHECK-LABEL: bitsel_v8i16:
; CHECK:       # %bb.0: # %entry
; CHECK-NEXT:    vbitsel.v $vr0, $vr0, $vr1, $vr2
; CHECK-NEXT:    ret
entry:
  %0 = and <8 x i16> %c, %b
  %1 = xor <8 x i16> %c, splat (i16 -1)
  %2 = and <8 x i16> %a, %1
  %3 = or <8 x i16> %0, %2
  ret <8 x i16> %3
}

define <4 x i32> @bitsel_v4i32(<4 x i32> %a, <4 x i32> %b, <4 x i32> %c) nounwind {
; CHECK-LABEL: bitsel_v4i32:
; CHECK:       # %bb.0: # %entry
; CHECK-NEXT:    vbitsel.v $vr0, $vr0, $vr1, $vr2
; CHECK-NEXT:    ret
entry:
  %0 = and <4 x i32> %c, %b
  %1 = xor <4 x i32> %c, splat (i32 -1)
  %2 = and <4 x i32> %a, %1
  %3 = or <4 x i32> %0, %2
  ret <4 x i32> %3
}

define <2 x i64> @bitsel_v2i64(<2 x i64> %a, <2 x i64> %b, <2 x i64> %c) nounwind {
; CHECK-LABEL: bitsel_v2i64:
; CHECK:       # %bb.0: # %entry
; CHECK-NEXT:    vbitsel.v $vr0, $vr0, $vr1, $vr2
; CHECK-NEXT:    ret
entry:
  %0 = and <2 x i64> %c, %b
  %1 = xor <2 x i64> %c, splat (i64 -1)
  %2 = and <2 x i64> %a, %1
  %3 = or <2 x i64> %0, %2
  ret <2 x i64> %3
}

define <16 x i8> @bitsel_v16i8_1(<16 x i8> %a, <16 x i8> %b) nounwind {
; CHECK-LABEL: bitsel_v16i8_1:
; CHECK:       # %bb.0: # %entry
; CHECK-NEXT:    vbitseli.b $vr0, $vr1, 1
; CHECK-NEXT:    ret
entry:
  %0 = and <16 x i8> %a, splat (i8 1)
  %1 = xor <16 x i8> %a, splat (i8 -1)
  %2 = and <16 x i8> %b, %1
  %3 = or <16 x i8> %0, %2
  ret <16 x i8> %3
}