; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=avx2 -stop-after=finalize-isel | FileCheck %s

define void @split_masked_store(ptr %0) {
  ; CHECK-LABEL: name: split_masked_store
  ; CHECK: bb.0.entry:
  ; CHECK-NEXT:   liveins: $rdi
  ; CHECK-NEXT: {{  $}}
  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:gr64 = COPY $rdi
  ; CHECK-NEXT:   [[AVX_SET0_:%[0-9]+]]:vr256 = AVX_SET0
  ; CHECK-NEXT:   [[VMOVAPSYrm:%[0-9]+]]:vr256 = VMOVAPSYrm $rip, 1, $noreg, %const.0, $noreg :: (load (s256) from constant-pool)
  ; CHECK-NEXT:   VMASKMOVPDYmr [[COPY]], 1, $noreg, 32, $noreg, killed [[VMOVAPSYrm]], [[AVX_SET0_]] :: (store unknown-size into %ir.0 + 32, align 8)
  ; CHECK-NEXT:   VMOVUPDYmr [[COPY]], 1, $noreg, 0, $noreg, [[AVX_SET0_]] :: (store (s256) into %ir.0, align 8)
  ; CHECK-NEXT:   RET 0
entry:
  call void @llvm.masked.store.v8f64.p0(<8 x double> zeroinitializer, ptr %0, i32 8, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 false, i1 false>)
  ret void
}

define <8 x double> @split_masked_load(ptr %0) {
  ; CHECK-LABEL: name: split_masked_load
  ; CHECK: bb.0.entry:
  ; CHECK-NEXT:   liveins: $rdi
  ; CHECK-NEXT: {{  $}}
  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:gr64 = COPY $rdi
  ; CHECK-NEXT:   [[VMOVAPSYrm:%[0-9]+]]:vr256 = VMOVAPSYrm $rip, 1, $noreg, %const.0, $noreg :: (load (s256) from constant-pool)
  ; CHECK-NEXT:   [[VMASKMOVPDYrm:%[0-9]+]]:vr256 = VMASKMOVPDYrm killed [[VMOVAPSYrm]], [[COPY]], 1, $noreg, 32, $noreg :: (load unknown-size from %ir.0 + 32, align 8)
  ; CHECK-NEXT:   [[VMOVUPDYrm:%[0-9]+]]:vr256 = VMOVUPDYrm [[COPY]], 1, $noreg, 0, $noreg :: (load (s256) from %ir.0, align 8)
  ; CHECK-NEXT:   $ymm0 = COPY [[VMOVUPDYrm]]
  ; CHECK-NEXT:   $ymm1 = COPY [[VMASKMOVPDYrm]]
  ; CHECK-NEXT:   RET 0, $ymm0, $ymm1
entry:
  %x = call <8 x double> @llvm.masked.load.v8f64.p0(ptr %0, i32 8, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 false, i1 false>, <8 x double> poison)
  ret <8 x double> %x
}

declare <8 x double> @llvm.masked.load.v8f64.p0(ptr, i32, <8 x i1>, <8 x double>)
declare void @llvm.masked.store.v8f64.p0(<8 x double>, ptr, i32, <8 x i1>)