; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -O0 < %s -mtriple=x86_64-linux-generic -verify-machineinstrs -mattr=sse2 | FileCheck --check-prefix=CHECK-O0 %s
; RUN: llc -O3 < %s -mtriple=x86_64-linux-generic -verify-machineinstrs -mattr=sse2 | FileCheck --check-prefix=CHECK-O3 %s

define i8 @load_i8(i8* %ptr) {
; CHECK-O0-LABEL: load_i8:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movb (%rdi), %al
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: load_i8:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movb (%rdi), %al
; CHECK-O3-NEXT:    retq
  %v = load atomic i8, i8* %ptr unordered, align 1
  ret i8 %v
}

define void @store_i8(i8* %ptr, i8 %v) {
; CHECK-O0-LABEL: store_i8:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movb %sil, %al
; CHECK-O0-NEXT:    movb %al, (%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: store_i8:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movb %sil, (%rdi)
; CHECK-O3-NEXT:    retq
  store atomic i8 %v, i8* %ptr unordered, align 1
  ret void
}

define i16 @load_i16(i16* %ptr) {
; CHECK-O0-LABEL: load_i16:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movw (%rdi), %ax
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: load_i16:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movzwl (%rdi), %eax
; CHECK-O3-NEXT:    retq
  %v = load atomic i16, i16* %ptr unordered, align 2
  ret i16 %v
}


define void @store_i16(i16* %ptr, i16 %v) {
; CHECK-O0-LABEL: store_i16:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movw %si, %ax
; CHECK-O0-NEXT:    movw %ax, (%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: store_i16:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movw %si, (%rdi)
; CHECK-O3-NEXT:    retq
  store atomic i16 %v, i16* %ptr unordered, align 2
  ret void
}

define i32 @load_i32(i32* %ptr) {
; CHECK-O0-LABEL: load_i32:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movl (%rdi), %eax
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: load_i32:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movl (%rdi), %eax
; CHECK-O3-NEXT:    retq
  %v = load atomic i32, i32* %ptr unordered, align 4
  ret i32 %v
}

define void @store_i32(i32* %ptr, i32 %v) {
; CHECK-O0-LABEL: store_i32:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movl %esi, (%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: store_i32:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movl %esi, (%rdi)
; CHECK-O3-NEXT:    retq
  store atomic i32 %v, i32* %ptr unordered, align 4
  ret void
}

define i64 @load_i64(i64* %ptr) {
; CHECK-O0-LABEL: load_i64:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movq (%rdi), %rax
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: load_i64:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movq (%rdi), %rax
; CHECK-O3-NEXT:    retq
  %v = load atomic i64, i64* %ptr unordered, align 8
  ret i64 %v
}

define void @store_i64(i64* %ptr, i64 %v) {
; CHECK-O0-LABEL: store_i64:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movq %rsi, (%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: store_i64:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movq %rsi, (%rdi)
; CHECK-O3-NEXT:    retq
  store atomic i64 %v, i64* %ptr unordered, align 8
  ret void
}

;; The next batch of tests are intended to show transforms which we
;; either *can't* do for legality, or don't currently implement.  The later
;; are noted carefully where relevant.

; Must use a full width op, not a byte op
define void @narrow_writeback_or(i64* %ptr) {
; CHECK-O0-LABEL: narrow_writeback_or:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movq (%rdi), %rax
; CHECK-O0-NEXT:    orq $7, %rax
; CHECK-O0-NEXT:    movq %rax, (%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: narrow_writeback_or:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    orq $7, (%rdi)
; CHECK-O3-NEXT:    retq
  %v = load atomic i64, i64* %ptr unordered, align 8
  %v.new = or i64 %v, 7
  store atomic i64 %v.new, i64* %ptr unordered, align 8
  ret void
}

; Must use a full width op, not a byte op
define void @narrow_writeback_and(i64* %ptr) {
; CHECK-O0-LABEL: narrow_writeback_and:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movq (%rdi), %rax
; CHECK-O0-NEXT:    movl %eax, %ecx
; CHECK-O0-NEXT:    andl $-256, %ecx
; CHECK-O0-NEXT:    movl %ecx, %eax
; CHECK-O0-NEXT:    movq %rax, (%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: narrow_writeback_and:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movl $4294967040, %eax # imm = 0xFFFFFF00
; CHECK-O3-NEXT:    andq %rax, (%rdi)
; CHECK-O3-NEXT:    retq
  %v = load atomic i64, i64* %ptr unordered, align 8
  %v.new = and i64 %v, 4294967040 ;; 0xFFFF_FF00
  store atomic i64 %v.new, i64* %ptr unordered, align 8
  ret void
}

; Must use a full width op, not a byte op
define void @narrow_writeback_xor(i64* %ptr) {
; CHECK-O0-LABEL: narrow_writeback_xor:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movq (%rdi), %rax
; CHECK-O0-NEXT:    xorq $7, %rax
; CHECK-O0-NEXT:    movq %rax, (%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: narrow_writeback_xor:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    xorq $7, (%rdi)
; CHECK-O3-NEXT:    retq
  %v = load atomic i64, i64* %ptr unordered, align 8
  %v.new = xor i64 %v, 7
  store atomic i64 %v.new, i64* %ptr unordered, align 8
  ret void
}

; Legal if wider type is also atomic (TODO)
define void @widen_store(i32* %p0, i32 %v1, i32 %v2) {
; CHECK-O0-LABEL: widen_store:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movl %esi, (%rdi)
; CHECK-O0-NEXT:    movl %edx, 4(%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: widen_store:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movl %esi, (%rdi)
; CHECK-O3-NEXT:    movl %edx, 4(%rdi)
; CHECK-O3-NEXT:    retq
  %p1 = getelementptr i32, i32* %p0, i64 1
  store atomic i32 %v1, i32* %p0 unordered, align 8
  store atomic i32 %v2, i32* %p1 unordered, align 4
  ret void
}

; Legal if wider type is also atomic (TODO)
define void @widen_broadcast(i32* %p0, i32 %v) {
; CHECK-O0-LABEL: widen_broadcast:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movl %esi, (%rdi)
; CHECK-O0-NEXT:    movl %esi, 4(%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: widen_broadcast:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movl %esi, (%rdi)
; CHECK-O3-NEXT:    movl %esi, 4(%rdi)
; CHECK-O3-NEXT:    retq
  %p1 = getelementptr i32, i32* %p0, i64 1
  store atomic i32 %v, i32* %p0 unordered, align 8
  store atomic i32 %v, i32* %p1 unordered, align 4
  ret void
}

; Legal if wider type is also atomic (TODO)
define void @vec_store(i32* %p0, <2 x i32> %vec) {
; CHECK-O0-LABEL: vec_store:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movd %xmm0, %eax
; CHECK-O0-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; CHECK-O0-NEXT:    movd %xmm0, %ecx
; CHECK-O0-NEXT:    movl %eax, (%rdi)
; CHECK-O0-NEXT:    movl %ecx, 4(%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: vec_store:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movd %xmm0, %eax
; CHECK-O3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
; CHECK-O3-NEXT:    movd %xmm0, %ecx
; CHECK-O3-NEXT:    movl %eax, (%rdi)
; CHECK-O3-NEXT:    movl %ecx, 4(%rdi)
; CHECK-O3-NEXT:    retq
  %v1 = extractelement <2 x i32> %vec, i32 0
  %v2 = extractelement <2 x i32> %vec, i32 1
  %p1 = getelementptr i32, i32* %p0, i64 1
  store atomic i32 %v1, i32* %p0 unordered, align 8
  store atomic i32 %v2, i32* %p1 unordered, align 4
  ret void
}


; Legal if wider type is also atomic (TODO)
; Also, can avoid register move from xmm to eax (TODO)
define void @widen_broadcast2(i32* %p0, <2 x i32> %vec) {
; CHECK-O0-LABEL: widen_broadcast2:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movd %xmm0, %eax
; CHECK-O0-NEXT:    movl %eax, (%rdi)
; CHECK-O0-NEXT:    movl %eax, 4(%rdi)
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: widen_broadcast2:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movd %xmm0, %eax
; CHECK-O3-NEXT:    movl %eax, (%rdi)
; CHECK-O3-NEXT:    movl %eax, 4(%rdi)
; CHECK-O3-NEXT:    retq
  %v1 = extractelement <2 x i32> %vec, i32 0
  %p1 = getelementptr i32, i32* %p0, i64 1
  store atomic i32 %v1, i32* %p0 unordered, align 8
  store atomic i32 %v1, i32* %p1 unordered, align 4
  ret void
}


; Legal if wider type is also atomic (TODO)
define void @widen_zero_init(i32* %p0, i32 %v1, i32 %v2) {
; CHECK-O0-LABEL: widen_zero_init:
; CHECK-O0:       # %bb.0:
; CHECK-O0-NEXT:    movl $0, (%rdi)
; CHECK-O0-NEXT:    movl $0, 4(%rdi)
; CHECK-O0-NEXT:    movl %esi, {{[-0-9]+}}(%r{{[sb]}}p) # 4-byte Spill
; CHECK-O0-NEXT:    movl %edx, {{[-0-9]+}}(%r{{[sb]}}p) # 4-byte Spill
; CHECK-O0-NEXT:    retq
;
; CHECK-O3-LABEL: widen_zero_init:
; CHECK-O3:       # %bb.0:
; CHECK-O3-NEXT:    movl $0, (%rdi)
; CHECK-O3-NEXT:    movl $0, 4(%rdi)
; CHECK-O3-NEXT:    retq
  %p1 = getelementptr i32, i32* %p0, i64 1
  store atomic i32 0, i32* %p0 unordered, align 8
  store atomic i32 0, i32* %p1 unordered, align 4
  ret void
}