llvm/test/Instrumentation/MemorySanitizer/expand-experimental-reductions.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt < %s -passes=msan -S | FileCheck %s
;
; Forked from llvm/test/CodeGen/Generic/expand-experimental-reductions.ll

target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"

declare i64 @llvm.vector.reduce.add.v2i64(<2 x i64>)
declare i64 @llvm.vector.reduce.mul.v2i64(<2 x i64>)
declare i64 @llvm.vector.reduce.and.v2i64(<2 x i64>)
declare i64 @llvm.vector.reduce.or.v2i64(<2 x i64>)
declare i64 @llvm.vector.reduce.xor.v2i64(<2 x i64>)

declare float @llvm.vector.reduce.fadd.f32.v4f32(float, <4 x float>)
declare float @llvm.vector.reduce.fmul.f32.v4f32(float, <4 x float>)

declare i64 @llvm.vector.reduce.smax.v2i64(<2 x i64>)
declare i64 @llvm.vector.reduce.smin.v2i64(<2 x i64>)
declare i64 @llvm.vector.reduce.umax.v2i64(<2 x i64>)
declare i64 @llvm.vector.reduce.umin.v2i64(<2 x i64>)

declare double @llvm.vector.reduce.fmax.v2f64(<2 x double>)
declare double @llvm.vector.reduce.fmin.v2f64(<2 x double>)

declare i8 @llvm.vector.reduce.and.i8.v3i8(<3 x i8>)

define i64 @add_i64(<2 x i64> %vec) #0 {
; CHECK-LABEL: define i64 @add_i64(
; CHECK-SAME: <2 x i64> [[VEC:%.*]]) #[[ATTR1:[0-9]+]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[R:%.*]] = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP1]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret i64 [[R]]
;
entry:
  %r = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %vec)
  ret i64 %r
}

define i64 @mul_i64(<2 x i64> %vec) #0 {
; CHECK-LABEL: define i64 @mul_i64(
; CHECK-SAME: <2 x i64> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[R:%.*]] = call i64 @llvm.vector.reduce.mul.v2i64(<2 x i64> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP1]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret i64 [[R]]
;
entry:
  %r = call i64 @llvm.vector.reduce.mul.v2i64(<2 x i64> %vec)
  ret i64 %r
}

define i64 @and_i64(<2 x i64> %vec) #0 {
; CHECK-LABEL: define i64 @and_i64(
; CHECK-SAME: <2 x i64> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = or <2 x i64> [[VEC]], [[TMP0]]
; CHECK-NEXT:    [[TMP2:%.*]] = call i64 @llvm.vector.reduce.and.v2i64(<2 x i64> [[TMP1]])
; CHECK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[TMP4:%.*]] = and i64 [[TMP2]], [[TMP3]]
; CHECK-NEXT:    [[R:%.*]] = call i64 @llvm.vector.reduce.and.v2i64(<2 x i64> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP4]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret i64 [[R]]
;
entry:
  %r = call i64 @llvm.vector.reduce.and.v2i64(<2 x i64> %vec)
  ret i64 %r
}

define i64 @or_i64(<2 x i64> %vec) #0 {
; CHECK-LABEL: define i64 @or_i64(
; CHECK-SAME: <2 x i64> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = xor <2 x i64> [[VEC]], splat (i64 -1)
; CHECK-NEXT:    [[TMP2:%.*]] = or <2 x i64> [[TMP1]], [[TMP0]]
; CHECK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vector.reduce.and.v2i64(<2 x i64> [[TMP2]])
; CHECK-NEXT:    [[TMP4:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i64 [[TMP3]], [[TMP4]]
; CHECK-NEXT:    [[R:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP5]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret i64 [[R]]
;
entry:
  %r = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> %vec)
  ret i64 %r
}

define i64 @xor_i64(<2 x i64> %vec) #0 {
; CHECK-LABEL: define i64 @xor_i64(
; CHECK-SAME: <2 x i64> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[R:%.*]] = call i64 @llvm.vector.reduce.xor.v2i64(<2 x i64> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP1]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret i64 [[R]]
;
entry:
  %r = call i64 @llvm.vector.reduce.xor.v2i64(<2 x i64> %vec)
  ret i64 %r
}

define float @fadd_f32(<4 x float> %vec) #0 {
; CHECK-LABEL: define float @fadd_f32(
; CHECK-SAME: <4 x float> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> [[TMP0]])
; CHECK-NEXT:    [[TMP2:%.*]] = or i32 0, [[TMP1]]
; CHECK-NEXT:    [[R:%.*]] = call fast float @llvm.vector.reduce.fadd.v4f32(float 0.000000e+00, <4 x float> [[VEC]])
; CHECK-NEXT:    store i32 [[TMP2]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret float [[R]]
;
entry:
  %r = call fast float @llvm.vector.reduce.fadd.f32.v4f32(float 0.0, <4 x float> %vec)
  ret float %r
}

define float @fadd_f32_accum(float %accum, <4 x float> %vec) #0 {
; CHECK-LABEL: define float @fadd_f32_accum(
; CHECK-SAME: float [[ACCUM:%.*]], <4 x float> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load i32, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP2:%.*]] = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> [[TMP1]])
; CHECK-NEXT:    [[TMP3:%.*]] = or i32 [[TMP0]], [[TMP2]]
; CHECK-NEXT:    [[R:%.*]] = call fast float @llvm.vector.reduce.fadd.v4f32(float [[ACCUM]], <4 x float> [[VEC]])
; CHECK-NEXT:    store i32 [[TMP3]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret float [[R]]
;
entry:
  %r = call fast float @llvm.vector.reduce.fadd.f32.v4f32(float %accum, <4 x float> %vec)
  ret float %r
}

define float @fadd_f32_strict(float %param, <4 x float> %vec) #0 {
; CHECK-LABEL: define float @fadd_f32_strict(
; CHECK-SAME: float [[PARAM:%.*]], <4 x float> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP3:%.*]] = load i32, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    [[TMP0:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> [[TMP0]])
; CHECK-NEXT:    [[TMP2:%.*]] = or i32 [[TMP3]], [[TMP1]]
; CHECK-NEXT:    [[R:%.*]] = call float @llvm.vector.reduce.fadd.v4f32(float [[PARAM]], <4 x float> [[VEC]])
; CHECK-NEXT:    store i32 [[TMP2]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret float [[R]]
;
entry:
  %r = call float @llvm.vector.reduce.fadd.f32.v4f32(float %param, <4 x float> %vec)
  ret float %r
}

define float @fadd_f32_strict_accum(float %accum, <4 x float> %vec) #0 {
; CHECK-LABEL: define float @fadd_f32_strict_accum(
; CHECK-SAME: float [[ACCUM:%.*]], <4 x float> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load i32, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP2:%.*]] = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> [[TMP1]])
; CHECK-NEXT:    [[TMP3:%.*]] = or i32 [[TMP0]], [[TMP2]]
; CHECK-NEXT:    [[R:%.*]] = call float @llvm.vector.reduce.fadd.v4f32(float [[ACCUM]], <4 x float> [[VEC]])
; CHECK-NEXT:    store i32 [[TMP3]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret float [[R]]
;
entry:
  %r = call float @llvm.vector.reduce.fadd.f32.v4f32(float %accum, <4 x float> %vec)
  ret float %r
}

define float @fmul_f32(<4 x float> %vec) #0 {
; CHECK-LABEL: define float @fmul_f32(
; CHECK-SAME: <4 x float> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> [[TMP0]])
; CHECK-NEXT:    [[TMP2:%.*]] = or i32 0, [[TMP1]]
; CHECK-NEXT:    [[R:%.*]] = call fast float @llvm.vector.reduce.fmul.v4f32(float 1.000000e+00, <4 x float> [[VEC]])
; CHECK-NEXT:    store i32 [[TMP2]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret float [[R]]
;
entry:
  %r = call fast float @llvm.vector.reduce.fmul.f32.v4f32(float 1.0, <4 x float> %vec)
  ret float %r
}

define float @fmul_f32_accum(float %accum, <4 x float> %vec) #0 {
; CHECK-LABEL: define float @fmul_f32_accum(
; CHECK-SAME: float [[ACCUM:%.*]], <4 x float> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load i32, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP2:%.*]] = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> [[TMP1]])
; CHECK-NEXT:    [[TMP3:%.*]] = or i32 [[TMP0]], [[TMP2]]
; CHECK-NEXT:    [[R:%.*]] = call fast float @llvm.vector.reduce.fmul.v4f32(float [[ACCUM]], <4 x float> [[VEC]])
; CHECK-NEXT:    store i32 [[TMP3]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret float [[R]]
;
entry:
  %r = call fast float @llvm.vector.reduce.fmul.f32.v4f32(float %accum, <4 x float> %vec)
  ret float %r
}

define float @fmul_f32_strict(float %param, <4 x float> %vec) #0 {
; CHECK-LABEL: define float @fmul_f32_strict(
; CHECK-SAME: float [[PARAM:%.*]], <4 x float> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP3:%.*]] = load i32, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    [[TMP0:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> [[TMP0]])
; CHECK-NEXT:    [[TMP2:%.*]] = or i32 [[TMP3]], [[TMP1]]
; CHECK-NEXT:    [[R:%.*]] = call float @llvm.vector.reduce.fmul.v4f32(float [[PARAM]], <4 x float> [[VEC]])
; CHECK-NEXT:    store i32 [[TMP2]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret float [[R]]
;
entry:
  %r = call float @llvm.vector.reduce.fmul.f32.v4f32(float %param, <4 x float> %vec)
  ret float %r
}

define float @fmul_f32_strict_accum(float %accum, <4 x float> %vec) #0 {
; CHECK-LABEL: define float @fmul_f32_strict_accum(
; CHECK-SAME: float [[ACCUM:%.*]], <4 x float> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load i32, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP2:%.*]] = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> [[TMP1]])
; CHECK-NEXT:    [[TMP3:%.*]] = or i32 [[TMP0]], [[TMP2]]
; CHECK-NEXT:    [[R:%.*]] = call float @llvm.vector.reduce.fmul.v4f32(float [[ACCUM]], <4 x float> [[VEC]])
; CHECK-NEXT:    store i32 [[TMP3]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret float [[R]]
;
entry:
  %r = call float @llvm.vector.reduce.fmul.f32.v4f32(float %accum, <4 x float> %vec)
  ret float %r
}

define i64 @smax_i64(<2 x i64> %vec) #0 {
; CHECK-LABEL: define i64 @smax_i64(
; CHECK-SAME: <2 x i64> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[R:%.*]] = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP1]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret i64 [[R]]
;
entry:
  %r = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> %vec)
  ret i64 %r
}

define i64 @smin_i64(<2 x i64> %vec) #0 {
; CHECK-LABEL: define i64 @smin_i64(
; CHECK-SAME: <2 x i64> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[R:%.*]] = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP1]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret i64 [[R]]
;
entry:
  %r = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> %vec)
  ret i64 %r
}

define i64 @umax_i64(<2 x i64> %vec) #0 {
; CHECK-LABEL: define i64 @umax_i64(
; CHECK-SAME: <2 x i64> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[R:%.*]] = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP1]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret i64 [[R]]
;
entry:
  %r = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> %vec)
  ret i64 %r
}

define i64 @umin_i64(<2 x i64> %vec) #0 {
; CHECK-LABEL: define i64 @umin_i64(
; CHECK-SAME: <2 x i64> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[R:%.*]] = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP1]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret i64 [[R]]
;
entry:
  %r = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> %vec)
  ret i64 %r
}


define double @fmax_f64(<2 x double> %vec) #0 {
; CHECK-LABEL: define double @fmax_f64(
; CHECK-SAME: <2 x double> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[R:%.*]] = call double @llvm.vector.reduce.fmax.v2f64(<2 x double> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP1]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret double [[R]]
;
entry:
  %r = call double @llvm.vector.reduce.fmax.v2f64(<2 x double> %vec)
  ret double %r
}


define double @fmin_f64(<2 x double> %vec) #0 {
; CHECK-LABEL: define double @fmin_f64(
; CHECK-SAME: <2 x double> [[VEC:%.*]]) #[[ATTR1]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> [[TMP0]])
; CHECK-NEXT:    [[R:%.*]] = call double @llvm.vector.reduce.fmin.v2f64(<2 x double> [[VEC]])
; CHECK-NEXT:    store i64 [[TMP1]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret double [[R]]
;
entry:
  %r = call double @llvm.vector.reduce.fmin.v2f64(<2 x double> %vec)
  ret double %r
}


define i8 @test_v3i8(<3 x i8> %a) nounwind #0 {
; CHECK-LABEL: define i8 @test_v3i8(
; CHECK-SAME: <3 x i8> [[A:%.*]]) #[[ATTR2:[0-9]+]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[TMP0:%.*]] = load <3 x i8>, ptr @__msan_param_tls, align 8
; CHECK-NEXT:    call void @llvm.donothing()
; CHECK-NEXT:    [[TMP1:%.*]] = or <3 x i8> [[A]], [[TMP0]]
; CHECK-NEXT:    [[TMP2:%.*]] = call i8 @llvm.vector.reduce.and.v3i8(<3 x i8> [[TMP1]])
; CHECK-NEXT:    [[TMP3:%.*]] = call i8 @llvm.vector.reduce.or.v3i8(<3 x i8> [[TMP0]])
; CHECK-NEXT:    [[TMP4:%.*]] = and i8 [[TMP2]], [[TMP3]]
; CHECK-NEXT:    [[B:%.*]] = call i8 @llvm.vector.reduce.and.v3i8(<3 x i8> [[A]])
; CHECK-NEXT:    store i8 [[TMP4]], ptr @__msan_retval_tls, align 8
; CHECK-NEXT:    ret i8 [[B]]
;
entry:
  %b = call i8 @llvm.vector.reduce.and.i8.v3i8(<3 x i8> %a)
  ret i8 %b
}

attributes #0 = { sanitize_memory }