; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgcn-amd-amdhsa < %s | FileCheck %s

; FIXME: Inefficient codegen which skips an optimization of load +
; extractelement when the vector element type is not byte-sized.
define i1 @extractloadi1(ptr %ptr, i32 %idx) {
; CHECK-LABEL: extractloadi1:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; CHECK-NEXT:    flat_load_ubyte v0, v[0:1]
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_lshrrev_b32_e32 v1, 2, v0
; CHECK-NEXT:    v_lshlrev_b32_e32 v3, 5, v0
; CHECK-NEXT:    v_and_b32_e32 v4, 2, v0
; CHECK-NEXT:    v_lshrrev_b32_e32 v5, 6, v0
; CHECK-NEXT:    v_lshrrev_b32_e32 v6, 4, v0
; CHECK-NEXT:    v_lshlrev_b32_e32 v7, 3, v0
; CHECK-NEXT:    v_lshlrev_b32_e32 v8, 1, v0
; CHECK-NEXT:    v_or_b32_e32 v1, v1, v3
; CHECK-NEXT:    v_and_b32_e32 v3, 0x100, v7
; CHECK-NEXT:    v_and_b32_e32 v7, 0x100, v8
; CHECK-NEXT:    v_lshlrev_b32_e32 v4, 7, v4
; CHECK-NEXT:    v_or_b32_e32 v3, v6, v3
; CHECK-NEXT:    v_or_b32_e32 v5, v5, v7
; CHECK-NEXT:    v_or_b32_e32 v0, v0, v4
; CHECK-NEXT:    v_and_b32_e32 v1, 0x103, v1
; CHECK-NEXT:    v_lshlrev_b32_e32 v4, 16, v5
; CHECK-NEXT:    v_lshlrev_b32_e32 v5, 16, v1
; CHECK-NEXT:    v_or_b32_e32 v1, v3, v4
; CHECK-NEXT:    v_or_b32_e32 v0, v0, v5
; CHECK-NEXT:    v_lshlrev_b32_e32 v2, 3, v2
; CHECK-NEXT:    v_lshr_b64 v[0:1], v[0:1], v2
; CHECK-NEXT:    s_setpc_b64 s[30:31]
  %val = load <8 x i1>, ptr %ptr
  %ret = extractelement <8 x i1> %val, i32 %idx
  ret i1 %ret
}