; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 3
; RUN: llc --mtriple=loongarch32 --mattr=+32s,+lasx < %s | FileCheck %s --check-prefixes=CHECK,LA32
; RUN: llc --mtriple=loongarch64 --mattr=+lasx < %s | FileCheck %s --check-prefixes=CHECK,LA64

define void @insert_32xi8(ptr %src, ptr %dst, i8 %in) nounwind {
; CHECK-LABEL: insert_32xi8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    xvld $xr0, $a0, 0
; CHECK-NEXT:    xvreplgr2vr.b $xr1, $a2
; CHECK-NEXT:    xvpermi.q $xr1, $xr0, 18
; CHECK-NEXT:    xvextrins.b $xr0, $xr1, 17
; CHECK-NEXT:    xvst $xr0, $a1, 0
; CHECK-NEXT:    ret
  %v = load volatile <32 x i8>, ptr %src
  %v_new = insertelement <32 x i8> %v, i8 %in, i32 1
  store <32 x i8> %v_new, ptr %dst
  ret void
}

define void @insert_32xi8_upper(ptr %src, ptr %dst, i8 %in) nounwind {
; CHECK-LABEL: insert_32xi8_upper:
; CHECK:       # %bb.0:
; CHECK-NEXT:    xvld $xr0, $a0, 0
; CHECK-NEXT:    xvreplgr2vr.b $xr1, $a2
; CHECK-NEXT:    xvpermi.q $xr1, $xr0, 48
; CHECK-NEXT:    xvextrins.b $xr0, $xr1, 0
; CHECK-NEXT:    xvst $xr0, $a1, 0
; CHECK-NEXT:    ret
  %v = load volatile <32 x i8>, ptr %src
  %v_new = insertelement <32 x i8> %v, i8 %in, i32 16
  store <32 x i8> %v_new, ptr %dst
  ret void
}

define void @insert_32xi8_undef(ptr %dst, i8 %in) nounwind {
; CHECK-LABEL: insert_32xi8_undef:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vinsgr2vr.b $vr0, $a1, 1
; CHECK-NEXT:    xvst $xr0, $a0, 0
; CHECK-NEXT:    ret
  %v = insertelement <32 x i8> poison, i8 %in, i32 1
  store <32 x i8> %v, ptr %dst
  ret void
}

define void @insert_32xi8_undef_upper(ptr %dst, i8 %in) nounwind {
; CHECK-LABEL: insert_32xi8_undef_upper:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vinsgr2vr.b $vr0, $a1, 6
; CHECK-NEXT:    xvpermi.q $xr0, $xr0, 2
; CHECK-NEXT:    xvst $xr0, $a0, 0
; CHECK-NEXT:    ret
  %v = insertelement <32 x i8> poison, i8 %in, i32 22
  store <32 x i8> %v, ptr %dst
  ret void
}

define void @insert_16xi16(ptr %src, ptr %dst, i16 %in) nounwind {
; CHECK-LABEL: insert_16xi16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    xvld $xr0, $a0, 0
; CHECK-NEXT:    xvreplgr2vr.h $xr1, $a2
; CHECK-NEXT:    xvpermi.q $xr1, $xr0, 18
; CHECK-NEXT:    xvextrins.h $xr0, $xr1, 17
; CHECK-NEXT:    xvst $xr0, $a1, 0
; CHECK-NEXT:    ret
  %v = load volatile <16 x i16>, ptr %src
  %v_new = insertelement <16 x i16> %v, i16 %in, i32 1
  store <16 x i16> %v_new, ptr %dst
  ret void
}

define void @insert_16xi16_upper(ptr %src, ptr %dst, i16 %in) nounwind {
; CHECK-LABEL: insert_16xi16_upper:
; CHECK:       # %bb.0:
; CHECK-NEXT:    xvld $xr0, $a0, 0
; CHECK-NEXT:    xvreplgr2vr.h $xr1, $a2
; CHECK-NEXT:    xvpermi.q $xr1, $xr0, 48
; CHECK-NEXT:    xvextrins.h $xr0, $xr1, 0
; CHECK-NEXT:    xvst $xr0, $a1, 0
; CHECK-NEXT:    ret
  %v = load volatile <16 x i16>, ptr %src
  %v_new = insertelement <16 x i16> %v, i16 %in, i32 8
  store <16 x i16> %v_new, ptr %dst
  ret void
}

define void @insert_16xi16_undef(ptr %dst, i16 %in) nounwind {
; CHECK-LABEL: insert_16xi16_undef:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vinsgr2vr.h $vr0, $a1, 1
; CHECK-NEXT:    xvst $xr0, $a0, 0
; CHECK-NEXT:    ret
  %v = insertelement <16 x i16> poison, i16 %in, i32 1
  store <16 x i16> %v, ptr %dst
  ret void
}

define void @insert_16xi16_undef_upper(ptr %dst, i16 %in) nounwind {
; CHECK-LABEL: insert_16xi16_undef_upper:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vinsgr2vr.h $vr0, $a1, 2
; CHECK-NEXT:    xvpermi.q $xr0, $xr0, 2
; CHECK-NEXT:    xvst $xr0, $a0, 0
; CHECK-NEXT:    ret
  %v = insertelement <16 x i16> poison, i16 %in, i32 10
  store <16 x i16> %v, ptr %dst
  ret void
}

define void @insert_8xi32(ptr %src, ptr %dst, i32 %in) nounwind {
; CHECK-LABEL: insert_8xi32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    xvld $xr0, $a0, 0
; CHECK-NEXT:    xvinsgr2vr.w $xr0, $a2, 1
; CHECK-NEXT:    xvst $xr0, $a1, 0
; CHECK-NEXT:    ret
  %v = load volatile <8 x i32>, ptr %src
  %v_new = insertelement <8 x i32> %v, i32 %in, i32 1
  store <8 x i32> %v_new, ptr %dst
  ret void
}

define void @insert_4xi64(ptr %src, ptr %dst, i64 %in) nounwind {
; LA32-LABEL: insert_4xi64:
; LA32:       # %bb.0:
; LA32-NEXT:    xvld $xr0, $a0, 0
; LA32-NEXT:    xvinsgr2vr.w $xr0, $a2, 2
; LA32-NEXT:    xvinsgr2vr.w $xr0, $a3, 3
; LA32-NEXT:    xvst $xr0, $a1, 0
; LA32-NEXT:    ret
;
; LA64-LABEL: insert_4xi64:
; LA64:       # %bb.0:
; LA64-NEXT:    xvld $xr0, $a0, 0
; LA64-NEXT:    xvinsgr2vr.d $xr0, $a2, 1
; LA64-NEXT:    xvst $xr0, $a1, 0
; LA64-NEXT:    ret
  %v = load volatile <4 x i64>, ptr %src
  %v_new = insertelement <4 x i64> %v, i64 %in, i32 1
  store <4 x i64> %v_new, ptr %dst
  ret void
}

define void @insert_8xfloat(ptr %src, ptr %dst, float %in) nounwind {
; CHECK-LABEL: insert_8xfloat:
; CHECK:       # %bb.0:
; CHECK-NEXT:    xvld $xr1, $a0, 0
; CHECK-NEXT:    # kill: def $f0 killed $f0 def $xr0
; CHECK-NEXT:    xvinsve0.w $xr1, $xr0, 1
; CHECK-NEXT:    xvst $xr1, $a1, 0
; CHECK-NEXT:    ret
  %v = load volatile <8 x float>, ptr %src
  %v_new = insertelement <8 x float> %v, float %in, i32 1
  store <8 x float> %v_new, ptr %dst
  ret void
}

define void @insert_4xdouble(ptr %src, ptr %dst, double %in) nounwind {
; CHECK-LABEL: insert_4xdouble:
; CHECK:       # %bb.0:
; CHECK-NEXT:    xvld $xr1, $a0, 0
; CHECK-NEXT:    # kill: def $f0_64 killed $f0_64 def $xr0
; CHECK-NEXT:    xvinsve0.d $xr1, $xr0, 1
; CHECK-NEXT:    xvst $xr1, $a1, 0
; CHECK-NEXT:    ret
  %v = load volatile <4 x double>, ptr %src
  %v_new = insertelement <4 x double> %v, double %in, i32 1
  store <4 x double> %v_new, ptr %dst
  ret void
}

define void @insert_32xi8_idx(ptr %src, ptr %dst, i8 %in, i32 %idx) nounwind {
; LA32-LABEL: insert_32xi8_idx:
; LA32:       # %bb.0:
; LA32-NEXT:    pcalau12i $a4, %pc_hi20(.LCPI12_0)
; LA32-NEXT:    xvld $xr0, $a4, %pc_lo12(.LCPI12_0)
; LA32-NEXT:    xvld $xr1, $a0, 0
; LA32-NEXT:    xvreplgr2vr.b $xr2, $a3
; LA32-NEXT:    xvseq.b $xr0, $xr2, $xr0
; LA32-NEXT:    xvreplgr2vr.b $xr2, $a2
; LA32-NEXT:    xvbitsel.v $xr0, $xr1, $xr2, $xr0
; LA32-NEXT:    xvst $xr0, $a1, 0
; LA32-NEXT:    ret
;
; LA64-LABEL: insert_32xi8_idx:
; LA64:       # %bb.0:
; LA64-NEXT:    pcalau12i $a4, %pc_hi20(.LCPI12_0)
; LA64-NEXT:    xvld $xr0, $a4, %pc_lo12(.LCPI12_0)
; LA64-NEXT:    xvld $xr1, $a0, 0
; LA64-NEXT:    bstrpick.d $a0, $a3, 31, 0
; LA64-NEXT:    xvreplgr2vr.b $xr2, $a0
; LA64-NEXT:    xvseq.b $xr0, $xr2, $xr0
; LA64-NEXT:    xvreplgr2vr.b $xr2, $a2
; LA64-NEXT:    xvbitsel.v $xr0, $xr1, $xr2, $xr0
; LA64-NEXT:    xvst $xr0, $a1, 0
; LA64-NEXT:    ret
  %v = load volatile <32 x i8>, ptr %src
  %v_new = insertelement <32 x i8> %v, i8 %in, i32 %idx
  store <32 x i8> %v_new, ptr %dst
  ret void
}

define void @insert_16xi16_idx(ptr %src, ptr %dst, i16 %in, i32 %idx) nounwind {
; LA32-LABEL: insert_16xi16_idx:
; LA32:       # %bb.0:
; LA32-NEXT:    pcalau12i $a4, %pc_hi20(.LCPI13_0)
; LA32-NEXT:    xvld $xr0, $a4, %pc_lo12(.LCPI13_0)
; LA32-NEXT:    xvld $xr1, $a0, 0
; LA32-NEXT:    xvreplgr2vr.h $xr2, $a3
; LA32-NEXT:    xvseq.h $xr0, $xr2, $xr0
; LA32-NEXT:    xvreplgr2vr.h $xr2, $a2
; LA32-NEXT:    xvbitsel.v $xr0, $xr1, $xr2, $xr0
; LA32-NEXT:    xvst $xr0, $a1, 0
; LA32-NEXT:    ret
;
; LA64-LABEL: insert_16xi16_idx:
; LA64:       # %bb.0:
; LA64-NEXT:    pcalau12i $a4, %pc_hi20(.LCPI13_0)
; LA64-NEXT:    xvld $xr0, $a4, %pc_lo12(.LCPI13_0)
; LA64-NEXT:    xvld $xr1, $a0, 0
; LA64-NEXT:    bstrpick.d $a0, $a3, 31, 0
; LA64-NEXT:    xvreplgr2vr.h $xr2, $a0
; LA64-NEXT:    xvseq.h $xr0, $xr2, $xr0
; LA64-NEXT:    xvreplgr2vr.h $xr2, $a2
; LA64-NEXT:    xvbitsel.v $xr0, $xr1, $xr2, $xr0
; LA64-NEXT:    xvst $xr0, $a1, 0
; LA64-NEXT:    ret
  %v = load volatile <16 x i16>, ptr %src
  %v_new = insertelement <16 x i16> %v, i16 %in, i32 %idx
  store <16 x i16> %v_new, ptr %dst
  ret void
}

define void @insert_8xi32_idx(ptr %src, ptr %dst, i32 %in, i32 %idx) nounwind {
; LA32-LABEL: insert_8xi32_idx:
; LA32:       # %bb.0:
; LA32-NEXT:    pcalau12i $a4, %pc_hi20(.LCPI14_0)
; LA32-NEXT:    xvld $xr0, $a4, %pc_lo12(.LCPI14_0)
; LA32-NEXT:    xvld $xr1, $a0, 0
; LA32-NEXT:    xvreplgr2vr.w $xr2, $a3
; LA32-NEXT:    xvseq.w $xr0, $xr2, $xr0
; LA32-NEXT:    xvreplgr2vr.w $xr2, $a2
; LA32-NEXT:    xvbitsel.v $xr0, $xr1, $xr2, $xr0
; LA32-NEXT:    xvst $xr0, $a1, 0
; LA32-NEXT:    ret
;
; LA64-LABEL: insert_8xi32_idx:
; LA64:       # %bb.0:
; LA64-NEXT:    pcalau12i $a4, %pc_hi20(.LCPI14_0)
; LA64-NEXT:    xvld $xr0, $a4, %pc_lo12(.LCPI14_0)
; LA64-NEXT:    xvld $xr1, $a0, 0
; LA64-NEXT:    bstrpick.d $a0, $a3, 31, 0
; LA64-NEXT:    xvreplgr2vr.w $xr2, $a0
; LA64-NEXT:    xvseq.w $xr0, $xr2, $xr0
; LA64-NEXT:    xvreplgr2vr.w $xr2, $a2
; LA64-NEXT:    xvbitsel.v $xr0, $xr1, $xr2, $xr0
; LA64-NEXT:    xvst $xr0, $a1, 0
; LA64-NEXT:    ret
  %v = load volatile <8 x i32>, ptr %src
  %v_new = insertelement <8 x i32> %v, i32 %in, i32 %idx
  store <8 x i32> %v_new, ptr %dst
  ret void
}

define void @insert_4xi64_idx(ptr %src, ptr %dst, i64 %in, i32 %idx) nounwind {
; LA32-LABEL: insert_4xi64_idx:
; LA32:       # %bb.0:
; LA32-NEXT:    pcalau12i $a5, %pc_hi20(.LCPI15_0)
; LA32-NEXT:    xvld $xr0, $a5, %pc_lo12(.LCPI15_0)
; LA32-NEXT:    add.w $a4, $a4, $a4
; LA32-NEXT:    xvld $xr1, $a0, 0
; LA32-NEXT:    xvreplgr2vr.w $xr2, $a4
; LA32-NEXT:    xvseq.w $xr2, $xr2, $xr0
; LA32-NEXT:    xvreplgr2vr.w $xr3, $a2
; LA32-NEXT:    xvbitsel.v $xr1, $xr1, $xr3, $xr2
; LA32-NEXT:    addi.w $a0, $a4, 1
; LA32-NEXT:    xvreplgr2vr.w $xr2, $a0
; LA32-NEXT:    xvseq.w $xr0, $xr2, $xr0
; LA32-NEXT:    xvreplgr2vr.w $xr2, $a3
; LA32-NEXT:    xvbitsel.v $xr0, $xr1, $xr2, $xr0
; LA32-NEXT:    xvst $xr0, $a1, 0
; LA32-NEXT:    ret
;
; LA64-LABEL: insert_4xi64_idx:
; LA64:       # %bb.0:
; LA64-NEXT:    pcalau12i $a4, %pc_hi20(.LCPI15_0)
; LA64-NEXT:    xvld $xr0, $a4, %pc_lo12(.LCPI15_0)
; LA64-NEXT:    xvld $xr1, $a0, 0
; LA64-NEXT:    bstrpick.d $a0, $a3, 31, 0
; LA64-NEXT:    xvreplgr2vr.d $xr2, $a0
; LA64-NEXT:    xvseq.d $xr0, $xr2, $xr0
; LA64-NEXT:    xvreplgr2vr.d $xr2, $a2
; LA64-NEXT:    xvbitsel.v $xr0, $xr1, $xr2, $xr0
; LA64-NEXT:    xvst $xr0, $a1, 0
; LA64-NEXT:    ret
  %v = load volatile <4 x i64>, ptr %src
  %v_new = insertelement <4 x i64> %v, i64 %in, i32 %idx
  store <4 x i64> %v_new, ptr %dst
  ret void
}

define void @insert_8xfloat_idx(ptr %src, ptr %dst, float %in, i32 %idx) nounwind {
; LA32-LABEL: insert_8xfloat_idx:
; LA32:       # %bb.0:
; LA32-NEXT:    pcalau12i $a3, %pc_hi20(.LCPI16_0)
; LA32-NEXT:    xvld $xr1, $a3, %pc_lo12(.LCPI16_0)
; LA32-NEXT:    # kill: def $f0 killed $f0 def $xr0
; LA32-NEXT:    xvld $xr2, $a0, 0
; LA32-NEXT:    xvreplgr2vr.w $xr3, $a2
; LA32-NEXT:    xvseq.w $xr1, $xr3, $xr1
; LA32-NEXT:    xvreplve0.w $xr0, $xr0
; LA32-NEXT:    xvbitsel.v $xr0, $xr2, $xr0, $xr1
; LA32-NEXT:    xvst $xr0, $a1, 0
; LA32-NEXT:    ret
;
; LA64-LABEL: insert_8xfloat_idx:
; LA64:       # %bb.0:
; LA64-NEXT:    # kill: def $f0 killed $f0 def $xr0
; LA64-NEXT:    pcalau12i $a3, %pc_hi20(.LCPI16_0)
; LA64-NEXT:    xvld $xr1, $a3, %pc_lo12(.LCPI16_0)
; LA64-NEXT:    xvld $xr2, $a0, 0
; LA64-NEXT:    bstrpick.d $a0, $a2, 31, 0
; LA64-NEXT:    xvreplgr2vr.w $xr3, $a0
; LA64-NEXT:    xvseq.w $xr1, $xr3, $xr1
; LA64-NEXT:    xvreplve0.w $xr0, $xr0
; LA64-NEXT:    xvbitsel.v $xr0, $xr2, $xr0, $xr1
; LA64-NEXT:    xvst $xr0, $a1, 0
; LA64-NEXT:    ret
  %v = load volatile <8 x float>, ptr %src
  %v_new = insertelement <8 x float> %v, float %in, i32 %idx
  store <8 x float> %v_new, ptr %dst
  ret void
}

define void @insert_4xdouble_idx(ptr %src, ptr %dst, double %in, i32 %idx) nounwind {
; LA32-LABEL: insert_4xdouble_idx:
; LA32:       # %bb.0:
; LA32-NEXT:    # kill: def $f0_64 killed $f0_64 def $xr0
; LA32-NEXT:    xvld $xr1, $a0, 0
; LA32-NEXT:    xvrepli.b $xr2, 0
; LA32-NEXT:    xvinsgr2vr.w $xr2, $a2, 0
; LA32-NEXT:    pcalau12i $a0, %pc_hi20(.LCPI17_0)
; LA32-NEXT:    xvld $xr3, $a0, %pc_lo12(.LCPI17_0)
; LA32-NEXT:    xvinsgr2vr.w $xr2, $a2, 2
; LA32-NEXT:    xvinsgr2vr.w $xr2, $a2, 4
; LA32-NEXT:    xvinsgr2vr.w $xr2, $a2, 6
; LA32-NEXT:    xvseq.d $xr2, $xr2, $xr3
; LA32-NEXT:    xvreplve0.d $xr0, $xr0
; LA32-NEXT:    xvbitsel.v $xr0, $xr1, $xr0, $xr2
; LA32-NEXT:    xvst $xr0, $a1, 0
; LA32-NEXT:    ret
;
; LA64-LABEL: insert_4xdouble_idx:
; LA64:       # %bb.0:
; LA64-NEXT:    # kill: def $f0_64 killed $f0_64 def $xr0
; LA64-NEXT:    pcalau12i $a3, %pc_hi20(.LCPI17_0)
; LA64-NEXT:    xvld $xr1, $a3, %pc_lo12(.LCPI17_0)
; LA64-NEXT:    xvld $xr2, $a0, 0
; LA64-NEXT:    bstrpick.d $a0, $a2, 31, 0
; LA64-NEXT:    xvreplgr2vr.d $xr3, $a0
; LA64-NEXT:    xvseq.d $xr1, $xr3, $xr1
; LA64-NEXT:    xvreplve0.d $xr0, $xr0
; LA64-NEXT:    xvbitsel.v $xr0, $xr2, $xr0, $xr1
; LA64-NEXT:    xvst $xr0, $a1, 0
; LA64-NEXT:    ret
  %v = load volatile <4 x double>, ptr %src
  %v_new = insertelement <4 x double> %v, double %in, i32 %idx
  store <4 x double> %v_new, ptr %dst
  ret void
}