llvm/test/CodeGen/AMDGPU/kill-true-in-return-block.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc -mtriple=amdgcn -mcpu=gfx90a %s -o - | FileCheck %s

define amdgpu_ps float @kill_true(i1 %.not) {
; CHECK-LABEL: kill_true:
; CHECK:       ; %bb.0: ; %entry
; CHECK-NEXT:    s_mov_b64 s[0:1], exec
; CHECK-NEXT:    s_wqm_b64 exec, exec
; CHECK-NEXT:    v_and_b32_e32 v0, 1, v0
; CHECK-NEXT:    v_cmp_ne_u32_e32 vcc, 1, v0
; CHECK-NEXT:    s_and_saveexec_b64 s[2:3], vcc
; CHECK-NEXT:    s_cbranch_execz .LBB0_2
; CHECK-NEXT:  ; %bb.1: ; %if1
; CHECK-NEXT:    s_mov_b32 s4, 0
; CHECK-NEXT:    ; kill: def $sgpr4 killed $sgpr4 killed $exec
; CHECK-NEXT:    v_pk_mov_b32 v[0:1], 0, 0
; CHECK-NEXT:    v_mov_b32_e32 v2, s4
; CHECK-NEXT:    flat_store_dword v[0:1], v2
; CHECK-NEXT:  .LBB0_2: ; %endif1
; CHECK-NEXT:    s_or_b64 exec, exec, s[2:3]
; CHECK-NEXT:    s_and_b64 exec, exec, s[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    ; return to shader part epilog
entry:
  br i1 %.not, label %endif1, label %if1

if1:
  %C = call float @llvm.amdgcn.wqm.f32(float 0.000000e+00)
  store float %C, ptr null, align 4
  br label %endif1

endif1:
  call void @llvm.amdgcn.kill(i1 true)
  ret float 0.000000e+00
}

declare void @llvm.amdgcn.kill(i1)

declare float @llvm.amdgcn.wqm.f32(float)