Index: test/CodeGen/AArch64/unfold-masked-merge-scalar.ll
===================================================================
--- /dev/null
+++ test/CodeGen/AArch64/unfold-masked-merge-scalar.ll
@@ -0,0 +1,722 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=aarch64-unknown-linux-gnu < %s | FileCheck %s
+
+; https://bugs.llvm.org/show_bug.cgi?id=37104
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 8-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i8 @out8(i8 %x, i8 %y, i8 %mask) {
+; CHECK-LABEL: out8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w0, w2
+; CHECK-NEXT:    bic w9, w1, w2
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %mx = and i8 %x, %mask
+  %notmask = xor i8 %mask, -1
+  %my = and i8 %y, %notmask
+  %r = or i8 %mx, %my
+  ret i8 %r
+}
+
+define i8 @out8_constmask(i8 %x, i8 %y) {
+; CHECK-LABEL: out8_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    lsr w8, w0, #2
+; CHECK-NEXT:    bfi w1, w8, #2, #4
+; CHECK-NEXT:    mov w0, w1
+; CHECK-NEXT:    ret
+  %mx = and i8 %x, 60
+  %my = and i8 %y, -61
+  %r = or i8 %mx, %my
+  ret i8 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 16-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i16 @out16(i16 %x, i16 %y, i16 %mask) {
+; CHECK-LABEL: out16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w0, w2
+; CHECK-NEXT:    bic w9, w1, w2
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %mx = and i16 %x, %mask
+  %notmask = xor i16 %mask, -1
+  %my = and i16 %y, %notmask
+  %r = or i16 %mx, %my
+  ret i16 %r
+}
+
+define i16 @out16_constmask(i16 %x, i16 %y) {
+; CHECK-LABEL: out16_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    lsr w8, w0, #4
+; CHECK-NEXT:    bfi w1, w8, #4, #8
+; CHECK-NEXT:    mov w0, w1
+; CHECK-NEXT:    ret
+  %mx = and i16 %x, 4080
+  %my = and i16 %y, -4081
+  %r = or i16 %mx, %my
+  ret i16 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 32-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i32 @out32(i32 %x, i32 %y, i32 %mask) {
+; CHECK-LABEL: out32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w0, w2
+; CHECK-NEXT:    bic w9, w1, w2
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %mx = and i32 %x, %mask
+  %notmask = xor i32 %mask, -1
+  %my = and i32 %y, %notmask
+  %r = or i32 %mx, %my
+  ret i32 %r
+}
+
+define i32 @out32_constmask(i32 %x, i32 %y) {
+; CHECK-LABEL: out32_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    lsr w8, w0, #8
+; CHECK-NEXT:    bfi w1, w8, #8, #16
+; CHECK-NEXT:    mov w0, w1
+; CHECK-NEXT:    ret
+  %mx = and i32 %x, 16776960
+  %my = and i32 %y, -16776961
+  %r = or i32 %mx, %my
+  ret i32 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 64-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i64 @out64(i64 %x, i64 %y, i64 %mask) {
+; CHECK-LABEL: out64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and x8, x0, x2
+; CHECK-NEXT:    bic x9, x1, x2
+; CHECK-NEXT:    orr x0, x8, x9
+; CHECK-NEXT:    ret
+  %mx = and i64 %x, %mask
+  %notmask = xor i64 %mask, -1
+  %my = and i64 %y, %notmask
+  %r = or i64 %mx, %my
+  ret i64 %r
+}
+
+define i64 @out64_constmask(i64 %x, i64 %y) {
+; CHECK-LABEL: out64_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    lsr x8, x0, #16
+; CHECK-NEXT:    bfi x1, x8, #16, #32
+; CHECK-NEXT:    mov x0, x1
+; CHECK-NEXT:    ret
+  %mx = and i64 %x, 281474976645120
+  %my = and i64 %y, -281474976645121
+  %r = or i64 %mx, %my
+  ret i64 %r
+}
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+; Should be the same as the previous one.
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 8-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i8 @in8(i8 %x, i8 %y, i8 %mask) {
+; CHECK-LABEL: in8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, w2
+; CHECK-NEXT:    eor w0, w8, w1
+; CHECK-NEXT:    ret
+  %n0 = xor i8 %x, %y
+  %n1 = and i8 %n0, %mask
+  %r = xor i8 %n1, %y
+  ret i8 %r
+}
+
+define i8 @in8_constmask(i8 %x, i8 %y) {
+; CHECK-LABEL: in8_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, #0x3c
+; CHECK-NEXT:    eor w0, w8, w1
+; CHECK-NEXT:    ret
+  %n0 = xor i8 %x, %y
+  %n1 = and i8 %n0, 60
+  %r = xor i8 %n1, %y
+  ret i8 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 16-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i16 @in16(i16 %x, i16 %y, i16 %mask) {
+; CHECK-LABEL: in16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, w2
+; CHECK-NEXT:    eor w0, w8, w1
+; CHECK-NEXT:    ret
+  %n0 = xor i16 %x, %y
+  %n1 = and i16 %n0, %mask
+  %r = xor i16 %n1, %y
+  ret i16 %r
+}
+
+define i16 @in16_constmask(i16 %x, i16 %y) {
+; CHECK-LABEL: in16_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, #0xff0
+; CHECK-NEXT:    eor w0, w8, w1
+; CHECK-NEXT:    ret
+  %n0 = xor i16 %x, %y
+  %n1 = and i16 %n0, 4080
+  %r = xor i16 %n1, %y
+  ret i16 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 32-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i32 @in32(i32 %x, i32 %y, i32 %mask) {
+; CHECK-LABEL: in32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, w2
+; CHECK-NEXT:    eor w0, w8, w1
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in32_constmask(i32 %x, i32 %y) {
+; CHECK-LABEL: in32_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, #0xffff00
+; CHECK-NEXT:    eor w0, w8, w1
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 64-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i64 @in64(i64 %x, i64 %y, i64 %mask) {
+; CHECK-LABEL: in64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor x8, x0, x1
+; CHECK-NEXT:    and x8, x8, x2
+; CHECK-NEXT:    eor x0, x8, x1
+; CHECK-NEXT:    ret
+  %n0 = xor i64 %x, %y
+  %n1 = and i64 %n0, %mask
+  %r = xor i64 %n1, %y
+  ret i64 %r
+}
+
+define i64 @in64_constmask(i64 %x, i64 %y) {
+; CHECK-LABEL: in64_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor x8, x0, x1
+; CHECK-NEXT:    and x8, x8, #0xffffffff0000
+; CHECK-NEXT:    eor x0, x8, x1
+; CHECK-NEXT:    ret
+  %n0 = xor i64 %x, %y
+  %n1 = and i64 %n0, 281474976645120
+  %r = xor i64 %n1, %y
+  ret i64 %r
+}
+
+; ============================================================================ ;
+; Commutativity tests.
+; ============================================================================ ;
+
+define i32 @in_commutativity_0_0_1(i32 %x, i32 %y, i32 %mask) {
+; CHECK-LABEL: in_commutativity_0_0_1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w2, w8
+; CHECK-NEXT:    eor w0, w8, w1
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0 ; swapped
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_commutativity_0_1_0(i32 %x, i32 %y, i32 %mask) {
+; CHECK-LABEL: in_commutativity_0_1_0:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, w2
+; CHECK-NEXT:    eor w0, w1, w8
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %y, %n1 ; swapped
+  ret i32 %r
+}
+
+define i32 @in_commutativity_0_1_1(i32 %x, i32 %y, i32 %mask) {
+; CHECK-LABEL: in_commutativity_0_1_1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w2, w8
+; CHECK-NEXT:    eor w0, w1, w8
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0 ; swapped
+  %r = xor i32 %y, %n1 ; swapped
+  ret i32 %r
+}
+
+define i32 @in_commutativity_1_0_0(i32 %x, i32 %y, i32 %mask) {
+; CHECK-LABEL: in_commutativity_1_0_0:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, w2
+; CHECK-NEXT:    eor w0, w8, w0
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %x ; %x instead of %y
+  ret i32 %r
+}
+
+define i32 @in_commutativity_1_0_1(i32 %x, i32 %y, i32 %mask) {
+; CHECK-LABEL: in_commutativity_1_0_1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w2, w8
+; CHECK-NEXT:    eor w0, w8, w0
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0 ; swapped
+  %r = xor i32 %n1, %x ; %x instead of %y
+  ret i32 %r
+}
+
+define i32 @in_commutativity_1_1_0(i32 %x, i32 %y, i32 %mask) {
+; CHECK-LABEL: in_commutativity_1_1_0:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, w2
+; CHECK-NEXT:    eor w0, w0, w8
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %x, %n1 ; swapped, %x instead of %y
+  ret i32 %r
+}
+
+define i32 @in_commutativity_1_1_1(i32 %x, i32 %y, i32 %mask) {
+; CHECK-LABEL: in_commutativity_1_1_1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w2, w8
+; CHECK-NEXT:    eor w0, w0, w8
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0 ; swapped
+  %r = xor i32 %x, %n1 ; swapped, %x instead of %y
+  ret i32 %r
+}
+
+; ============================================================================ ;
+; Constant Commutativity tests.
+; ============================================================================ ;
+
+define i32 @in_constmask_commutativity_0_1(i32 %x, i32 %y) {
+; CHECK-LABEL: in_constmask_commutativity_0_1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, #0xffff00
+; CHECK-NEXT:    eor w0, w1, w8
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %y, %n1 ; swapped
+  ret i32 %r
+}
+
+define i32 @in_constmask_commutativity_1_0(i32 %x, i32 %y) {
+; CHECK-LABEL: in_constmask_commutativity_1_0:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, #0xffff00
+; CHECK-NEXT:    eor w0, w8, w0
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %n1, %x ; %x instead of %y
+  ret i32 %r
+}
+
+define i32 @in_constmask_commutativity_1_1(i32 %x, i32 %y) {
+; CHECK-LABEL: in_constmask_commutativity_1_1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, #0xffff00
+; CHECK-NEXT:    eor w0, w0, w8
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %x, %n1 ; swapped, %x instead of %y
+  ret i32 %r
+}
+
+; ============================================================================ ;
+; Y is an 'and' too.
+; ============================================================================ ;
+
+define i32 @in_complex_y0(i32 %x, i32 %y_hi, i32 %y_low, i32 %mask) {
+; CHECK-LABEL: in_complex_y0:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w1, w2
+; CHECK-NEXT:    eor w9, w0, w8
+; CHECK-NEXT:    and w9, w9, w3
+; CHECK-NEXT:    eor w0, w9, w8
+; CHECK-NEXT:    ret
+  %y = and i32 %y_hi, %y_low
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_complex_y1(i32 %x, i32 %y_hi, i32 %y_low, i32 %mask) {
+; CHECK-LABEL: in_complex_y1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w1, w2
+; CHECK-NEXT:    eor w9, w0, w8
+; CHECK-NEXT:    and w9, w9, w3
+; CHECK-NEXT:    eor w0, w8, w9
+; CHECK-NEXT:    ret
+  %y = and i32 %y_hi, %y_low
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %y, %n1
+  ret i32 %r
+}
+
+define i32 @in_complex_y0_constmask(i32 %x, i32 %y_hi, i32 %y_low) {
+; CHECK-LABEL: in_complex_y0_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w1, w2
+; CHECK-NEXT:    eor w9, w0, w8
+; CHECK-NEXT:    and w9, w9, #0xffff00
+; CHECK-NEXT:    eor w0, w9, w8
+; CHECK-NEXT:    ret
+  %y = and i32 %y_hi, %y_low
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_complex_y1_constmask(i32 %x, i32 %y_hi, i32 %y_low) {
+; CHECK-LABEL: in_complex_y1_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w1, w2
+; CHECK-NEXT:    eor w9, w0, w8
+; CHECK-NEXT:    and w9, w9, #0xffff00
+; CHECK-NEXT:    eor w0, w8, w9
+; CHECK-NEXT:    ret
+  %y = and i32 %y_hi, %y_low
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %y, %n1
+  ret i32 %r
+}
+
+; ============================================================================ ;
+; M is an 'xor' too.
+; ============================================================================ ;
+
+define i32 @in_complex_m0(i32 %x, i32 %y, i32 %m_a, i32 %m_b) {
+; CHECK-LABEL: in_complex_m0:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w2, w3
+; CHECK-NEXT:    eor w9, w0, w1
+; CHECK-NEXT:    and w8, w9, w8
+; CHECK-NEXT:    eor w0, w8, w1
+; CHECK-NEXT:    ret
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_complex_m1(i32 %x, i32 %y, i32 %m_a, i32 %m_b) {
+; CHECK-LABEL: in_complex_m1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w2, w3
+; CHECK-NEXT:    eor w9, w0, w1
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    eor w0, w8, w1
+; CHECK-NEXT:    ret
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+; ============================================================================ ;
+; Both Y and M are complex.
+; ============================================================================ ;
+
+define i32 @in_complex_y0_m0(i32 %x, i32 %y_hi, i32 %y_low, i32 %m_a, i32 %m_b) {
+; CHECK-LABEL: in_complex_y0_m0:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w1, w2
+; CHECK-NEXT:    eor w9, w3, w4
+; CHECK-NEXT:    eor w10, w0, w8
+; CHECK-NEXT:    and w9, w10, w9
+; CHECK-NEXT:    eor w0, w9, w8
+; CHECK-NEXT:    ret
+  %y = and i32 %y_hi, %y_low
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_complex_y1_m0(i32 %x, i32 %y_hi, i32 %y_low, i32 %m_a, i32 %m_b) {
+; CHECK-LABEL: in_complex_y1_m0:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w1, w2
+; CHECK-NEXT:    eor w9, w3, w4
+; CHECK-NEXT:    eor w10, w0, w8
+; CHECK-NEXT:    and w9, w10, w9
+; CHECK-NEXT:    eor w0, w8, w9
+; CHECK-NEXT:    ret
+  %y = and i32 %y_hi, %y_low
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %y, %n1
+  ret i32 %r
+}
+
+define i32 @in_complex_y0_m1(i32 %x, i32 %y_hi, i32 %y_low, i32 %m_a, i32 %m_b) {
+; CHECK-LABEL: in_complex_y0_m1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w1, w2
+; CHECK-NEXT:    eor w9, w3, w4
+; CHECK-NEXT:    eor w10, w0, w8
+; CHECK-NEXT:    and w9, w9, w10
+; CHECK-NEXT:    eor w0, w9, w8
+; CHECK-NEXT:    ret
+  %y = and i32 %y_hi, %y_low
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_complex_y1_m1(i32 %x, i32 %y_hi, i32 %y_low, i32 %m_a, i32 %m_b) {
+; CHECK-LABEL: in_complex_y1_m1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w1, w2
+; CHECK-NEXT:    eor w9, w3, w4
+; CHECK-NEXT:    eor w10, w0, w8
+; CHECK-NEXT:    and w9, w9, w10
+; CHECK-NEXT:    eor w0, w8, w9
+; CHECK-NEXT:    ret
+  %y = and i32 %y_hi, %y_low
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0
+  %r = xor i32 %y, %n1
+  ret i32 %r
+}
+
+; ============================================================================ ;
+; Negative tests. Should not be folded.
+; ============================================================================ ;
+
+; Multi-use tests.
+
+declare void @use32(i32)
+
+define i32 @in_multiuse_A(i32 %x, i32 %y, i32 %z, i32 %mask) {
+; CHECK-LABEL: in_multiuse_A:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    str x20, [sp, #-32]! // 8-byte Folded Spill
+; CHECK-NEXT:    stp x19, x30, [sp, #16] // 8-byte Folded Spill
+; CHECK-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-NEXT:    .cfi_offset w30, -8
+; CHECK-NEXT:    .cfi_offset w19, -16
+; CHECK-NEXT:    .cfi_offset w20, -32
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w20, w8, w3
+; CHECK-NEXT:    mov w0, w20
+; CHECK-NEXT:    mov w19, w1
+; CHECK-NEXT:    bl use32
+; CHECK-NEXT:    eor w0, w20, w19
+; CHECK-NEXT:    ldp x19, x30, [sp, #16] // 8-byte Folded Reload
+; CHECK-NEXT:    ldr x20, [sp], #32 // 8-byte Folded Reload
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  call void @use32(i32 %n1)
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_multiuse_A_constmask(i32 %x, i32 %y, i32 %z) {
+; CHECK-LABEL: in_multiuse_A_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    str x20, [sp, #-32]! // 8-byte Folded Spill
+; CHECK-NEXT:    stp x19, x30, [sp, #16] // 8-byte Folded Spill
+; CHECK-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-NEXT:    .cfi_offset w30, -8
+; CHECK-NEXT:    .cfi_offset w19, -16
+; CHECK-NEXT:    .cfi_offset w20, -32
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w20, w8, #0xffff00
+; CHECK-NEXT:    mov w0, w20
+; CHECK-NEXT:    mov w19, w1
+; CHECK-NEXT:    bl use32
+; CHECK-NEXT:    eor w0, w20, w19
+; CHECK-NEXT:    ldp x19, x30, [sp, #16] // 8-byte Folded Reload
+; CHECK-NEXT:    ldr x20, [sp], #32 // 8-byte Folded Reload
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  call void @use32(i32 %n1)
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_multiuse_B(i32 %x, i32 %y, i32 %z, i32 %mask) {
+; CHECK-LABEL: in_multiuse_B:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    str x20, [sp, #-32]! // 8-byte Folded Spill
+; CHECK-NEXT:    stp x19, x30, [sp, #16] // 8-byte Folded Spill
+; CHECK-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-NEXT:    .cfi_offset w30, -8
+; CHECK-NEXT:    .cfi_offset w19, -16
+; CHECK-NEXT:    .cfi_offset w20, -32
+; CHECK-NEXT:    eor w0, w0, w1
+; CHECK-NEXT:    mov w19, w1
+; CHECK-NEXT:    and w20, w0, w3
+; CHECK-NEXT:    bl use32
+; CHECK-NEXT:    eor w0, w20, w19
+; CHECK-NEXT:    ldp x19, x30, [sp, #16] // 8-byte Folded Reload
+; CHECK-NEXT:    ldr x20, [sp], #32 // 8-byte Folded Reload
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  call void @use32(i32 %n0)
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_multiuse_B_constmask(i32 %x, i32 %y, i32 %z) {
+; CHECK-LABEL: in_multiuse_B_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    str x20, [sp, #-32]! // 8-byte Folded Spill
+; CHECK-NEXT:    stp x19, x30, [sp, #16] // 8-byte Folded Spill
+; CHECK-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-NEXT:    .cfi_offset w30, -8
+; CHECK-NEXT:    .cfi_offset w19, -16
+; CHECK-NEXT:    .cfi_offset w20, -32
+; CHECK-NEXT:    eor w0, w0, w1
+; CHECK-NEXT:    mov w19, w1
+; CHECK-NEXT:    and w20, w0, #0xffff00
+; CHECK-NEXT:    bl use32
+; CHECK-NEXT:    eor w0, w20, w19
+; CHECK-NEXT:    ldp x19, x30, [sp, #16] // 8-byte Folded Reload
+; CHECK-NEXT:    ldr x20, [sp], #32 // 8-byte Folded Reload
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  call void @use32(i32 %n0)
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+; Various bad variants
+
+define i32 @n0_badmask(i32 %x, i32 %y, i32 %mask, i32 %mask2) {
+; CHECK-LABEL: n0_badmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w0, w2
+; CHECK-NEXT:    bic w9, w1, w3
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %mx = and i32 %x, %mask
+  %notmask = xor i32 %mask2, -1 ; %mask2 instead of %mask
+  %my = and i32 %y, %notmask
+  %r = or i32 %mx, %my
+  ret i32 %r
+}
+
+define i32 @n0_badxor(i32 %x, i32 %y, i32 %mask) {
+; CHECK-LABEL: n0_badxor:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w9, w2, #0x1
+; CHECK-NEXT:    and w8, w0, w2
+; CHECK-NEXT:    and w9, w1, w9
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %mx = and i32 %x, %mask
+  %notmask = xor i32 %mask, 1 ; instead of -1
+  %my = and i32 %y, %notmask
+  %r = or i32 %mx, %my
+  ret i32 %r
+}
+
+define i32 @n0_badconstmask(i32 %x, i32 %y) {
+; CHECK-LABEL: n0_badconstmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    mov w9, #256
+; CHECK-NEXT:    movk w9, #65280, lsl #16
+; CHECK-NEXT:    and w8, w0, #0xffff00
+; CHECK-NEXT:    and w9, w1, w9
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %mx = and i32 %x, 16776960
+  %my = and i32 %y, -16776960 ; instead of -16776961
+  %r = or i32 %mx, %my
+  ret i32 %r
+}
+
+define i32 @n1_thirdvar(i32 %x, i32 %y, i32 %z, i32 %mask) {
+; CHECK-LABEL: n1_thirdvar:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, w3
+; CHECK-NEXT:    eor w0, w8, w2
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %z ; instead of %y
+  ret i32 %r
+}
+
+define i32 @n1_thirdvar_constmask(i32 %x, i32 %y, i32 %z) {
+; CHECK-LABEL: n1_thirdvar_constmask:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    and w8, w8, #0xffff00
+; CHECK-NEXT:    eor w0, w8, w2
+; CHECK-NEXT:    ret
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %n1, %z ; instead of %y
+  ret i32 %r
+}
Index: test/CodeGen/X86/unfold-masked-merge-scalar.ll
===================================================================
--- /dev/null
+++ test/CodeGen/X86/unfold-masked-merge-scalar.ll
@@ -0,0 +1,1226 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mattr=-bmi < %s | FileCheck %s --check-prefix=CHECK-NOBMI
+; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mattr=+bmi < %s | FileCheck %s --check-prefix=CHECK-BMI
+
+; https://bugs.llvm.org/show_bug.cgi?id=37104
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 8-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i8 @out8(i8 %x, i8 %y, i8 %mask) {
+; CHECK-NOBMI-LABEL: out8:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    notb %dl
+; CHECK-NOBMI-NEXT:    andb %sil, %dl
+; CHECK-NOBMI-NEXT:    orb %dil, %dl
+; CHECK-NOBMI-NEXT:    movl %edx, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: out8:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    notb %dl
+; CHECK-BMI-NEXT:    andb %sil, %dl
+; CHECK-BMI-NEXT:    orb %dil, %dl
+; CHECK-BMI-NEXT:    movl %edx, %eax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i8 %x, %mask
+  %notmask = xor i8 %mask, -1
+  %my = and i8 %y, %notmask
+  %r = or i8 %mx, %my
+  ret i8 %r
+}
+
+define i8 @out8_constmask(i8 %x, i8 %y) {
+; CHECK-NOBMI-LABEL: out8_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andb $60, %dil
+; CHECK-NOBMI-NEXT:    andb $-61, %sil
+; CHECK-NOBMI-NEXT:    orb %dil, %sil
+; CHECK-NOBMI-NEXT:    movl %esi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: out8_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andb $60, %dil
+; CHECK-BMI-NEXT:    andb $-61, %sil
+; CHECK-BMI-NEXT:    orb %dil, %sil
+; CHECK-BMI-NEXT:    movl %esi, %eax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i8 %x, 60
+  %my = and i8 %y, -61
+  %r = or i8 %mx, %my
+  ret i8 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 16-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i16 @out16(i16 %x, i16 %y, i16 %mask) {
+; CHECK-NOBMI-LABEL: out16:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    notl %edx
+; CHECK-NOBMI-NEXT:    andl %esi, %edx
+; CHECK-NOBMI-NEXT:    orl %edi, %edx
+; CHECK-NOBMI-NEXT:    movl %edx, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: out16:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    andnl %esi, %edx, %eax
+; CHECK-BMI-NEXT:    orl %edi, %eax
+; CHECK-BMI-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i16 %x, %mask
+  %notmask = xor i16 %mask, -1
+  %my = and i16 %y, %notmask
+  %r = or i16 %mx, %my
+  ret i16 %r
+}
+
+define i16 @out16_constmask(i16 %x, i16 %y) {
+; CHECK-NOBMI-LABEL: out16_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    # kill: def $esi killed $esi def $rsi
+; CHECK-NOBMI-NEXT:    # kill: def $edi killed $edi def $rdi
+; CHECK-NOBMI-NEXT:    andl $4080, %edi # imm = 0xFF0
+; CHECK-NOBMI-NEXT:    andl $-4081, %esi # imm = 0xF00F
+; CHECK-NOBMI-NEXT:    leal (%rsi,%rdi), %eax
+; CHECK-NOBMI-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: out16_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    # kill: def $esi killed $esi def $rsi
+; CHECK-BMI-NEXT:    # kill: def $edi killed $edi def $rdi
+; CHECK-BMI-NEXT:    andl $4080, %edi # imm = 0xFF0
+; CHECK-BMI-NEXT:    andl $-4081, %esi # imm = 0xF00F
+; CHECK-BMI-NEXT:    leal (%rsi,%rdi), %eax
+; CHECK-BMI-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i16 %x, 4080
+  %my = and i16 %y, -4081
+  %r = or i16 %mx, %my
+  ret i16 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 32-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i32 @out32(i32 %x, i32 %y, i32 %mask) {
+; CHECK-NOBMI-LABEL: out32:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    notl %edx
+; CHECK-NOBMI-NEXT:    andl %esi, %edx
+; CHECK-NOBMI-NEXT:    orl %edi, %edx
+; CHECK-NOBMI-NEXT:    movl %edx, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: out32:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    andnl %esi, %edx, %eax
+; CHECK-BMI-NEXT:    orl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i32 %x, %mask
+  %notmask = xor i32 %mask, -1
+  %my = and i32 %y, %notmask
+  %r = or i32 %mx, %my
+  ret i32 %r
+}
+
+define i32 @out32_constmask(i32 %x, i32 %y) {
+; CHECK-NOBMI-LABEL: out32_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    # kill: def $esi killed $esi def $rsi
+; CHECK-NOBMI-NEXT:    # kill: def $edi killed $edi def $rdi
+; CHECK-NOBMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    andl $-16776961, %esi # imm = 0xFF0000FF
+; CHECK-NOBMI-NEXT:    leal (%rsi,%rdi), %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: out32_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    # kill: def $esi killed $esi def $rsi
+; CHECK-BMI-NEXT:    # kill: def $edi killed $edi def $rdi
+; CHECK-BMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    andl $-16776961, %esi # imm = 0xFF0000FF
+; CHECK-BMI-NEXT:    leal (%rsi,%rdi), %eax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i32 %x, 16776960
+  %my = and i32 %y, -16776961
+  %r = or i32 %mx, %my
+  ret i32 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 64-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i64 @out64(i64 %x, i64 %y, i64 %mask) {
+; CHECK-NOBMI-LABEL: out64:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andq %rdx, %rdi
+; CHECK-NOBMI-NEXT:    notq %rdx
+; CHECK-NOBMI-NEXT:    andq %rsi, %rdx
+; CHECK-NOBMI-NEXT:    orq %rdi, %rdx
+; CHECK-NOBMI-NEXT:    movq %rdx, %rax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: out64:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andq %rdx, %rdi
+; CHECK-BMI-NEXT:    andnq %rsi, %rdx, %rax
+; CHECK-BMI-NEXT:    orq %rdi, %rax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i64 %x, %mask
+  %notmask = xor i64 %mask, -1
+  %my = and i64 %y, %notmask
+  %r = or i64 %mx, %my
+  ret i64 %r
+}
+
+define i64 @out64_constmask(i64 %x, i64 %y) {
+; CHECK-NOBMI-LABEL: out64_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    movabsq $281474976645120, %rcx # imm = 0xFFFFFFFF0000
+; CHECK-NOBMI-NEXT:    andq %rdi, %rcx
+; CHECK-NOBMI-NEXT:    movabsq $-281474976645121, %rax # imm = 0xFFFF00000000FFFF
+; CHECK-NOBMI-NEXT:    andq %rsi, %rax
+; CHECK-NOBMI-NEXT:    orq %rcx, %rax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: out64_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    movabsq $281474976645120, %rcx # imm = 0xFFFFFFFF0000
+; CHECK-BMI-NEXT:    andq %rdi, %rcx
+; CHECK-BMI-NEXT:    movabsq $-281474976645121, %rax # imm = 0xFFFF00000000FFFF
+; CHECK-BMI-NEXT:    andq %rsi, %rax
+; CHECK-BMI-NEXT:    orq %rcx, %rax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i64 %x, 281474976645120
+  %my = and i64 %y, -281474976645121
+  %r = or i64 %mx, %my
+  ret i64 %r
+}
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+; Should be the same as the previous one.
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 8-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i8 @in8(i8 %x, i8 %y, i8 %mask) {
+; CHECK-NOBMI-LABEL: in8:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in8:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i8 %x, %y
+  %n1 = and i8 %n0, %mask
+  %r = xor i8 %n1, %y
+  ret i8 %r
+}
+
+define i8 @in8_constmask(i8 %x, i8 %y) {
+; CHECK-NOBMI-LABEL: in8_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andb $60, %dil
+; CHECK-NOBMI-NEXT:    xorb %dil, %sil
+; CHECK-NOBMI-NEXT:    movl %esi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in8_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andb $60, %dil
+; CHECK-BMI-NEXT:    xorb %dil, %sil
+; CHECK-BMI-NEXT:    movl %esi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i8 %x, %y
+  %n1 = and i8 %n0, 60
+  %r = xor i8 %n1, %y
+  ret i8 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 16-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i16 @in16(i16 %x, i16 %y, i16 %mask) {
+; CHECK-NOBMI-LABEL: in16:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in16:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i16 %x, %y
+  %n1 = and i16 %n0, %mask
+  %r = xor i16 %n1, %y
+  ret i16 %r
+}
+
+define i16 @in16_constmask(i16 %x, i16 %y) {
+; CHECK-NOBMI-LABEL: in16_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl $4080, %edi # imm = 0xFF0
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in16_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl $4080, %edi # imm = 0xFF0
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i16 %x, %y
+  %n1 = and i16 %n0, 4080
+  %r = xor i16 %n1, %y
+  ret i16 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 32-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i32 @in32(i32 %x, i32 %y, i32 %mask) {
+; CHECK-NOBMI-LABEL: in32:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in32:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in32_constmask(i32 %x, i32 %y) {
+; CHECK-NOBMI-LABEL: in32_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in32_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 64-bit ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+define i64 @in64(i64 %x, i64 %y, i64 %mask) {
+; CHECK-NOBMI-LABEL: in64:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorq %rsi, %rdi
+; CHECK-NOBMI-NEXT:    andq %rdx, %rdi
+; CHECK-NOBMI-NEXT:    xorq %rsi, %rdi
+; CHECK-NOBMI-NEXT:    movq %rdi, %rax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in64:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorq %rsi, %rdi
+; CHECK-BMI-NEXT:    andq %rdx, %rdi
+; CHECK-BMI-NEXT:    xorq %rsi, %rdi
+; CHECK-BMI-NEXT:    movq %rdi, %rax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i64 %x, %y
+  %n1 = and i64 %n0, %mask
+  %r = xor i64 %n1, %y
+  ret i64 %r
+}
+
+define i64 @in64_constmask(i64 %x, i64 %y) {
+; CHECK-NOBMI-LABEL: in64_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorq %rsi, %rdi
+; CHECK-NOBMI-NEXT:    movabsq $281474976645120, %rax # imm = 0xFFFFFFFF0000
+; CHECK-NOBMI-NEXT:    andq %rdi, %rax
+; CHECK-NOBMI-NEXT:    xorq %rsi, %rax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in64_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorq %rsi, %rdi
+; CHECK-BMI-NEXT:    movabsq $281474976645120, %rax # imm = 0xFFFFFFFF0000
+; CHECK-BMI-NEXT:    andq %rdi, %rax
+; CHECK-BMI-NEXT:    xorq %rsi, %rax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i64 %x, %y
+  %n1 = and i64 %n0, 281474976645120
+  %r = xor i64 %n1, %y
+  ret i64 %r
+}
+
+; ============================================================================ ;
+; Commutativity tests.
+; ============================================================================ ;
+
+define i32 @in_commutativity_0_0_1(i32 %x, i32 %y, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_commutativity_0_0_1:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_commutativity_0_0_1:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0 ; swapped
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_commutativity_0_1_0(i32 %x, i32 %y, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_commutativity_0_1_0:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_commutativity_0_1_0:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %y, %n1 ; swapped
+  ret i32 %r
+}
+
+define i32 @in_commutativity_0_1_1(i32 %x, i32 %y, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_commutativity_0_1_1:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_commutativity_0_1_1:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0 ; swapped
+  %r = xor i32 %y, %n1 ; swapped
+  ret i32 %r
+}
+
+define i32 @in_commutativity_1_0_0(i32 %x, i32 %y, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_commutativity_1_0_0:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    movl %esi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_commutativity_1_0_0:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    movl %esi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %x ; %x instead of %y
+  ret i32 %r
+}
+
+define i32 @in_commutativity_1_0_1(i32 %x, i32 %y, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_commutativity_1_0_1:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    movl %esi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_commutativity_1_0_1:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    movl %esi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0 ; swapped
+  %r = xor i32 %n1, %x ; %x instead of %y
+  ret i32 %r
+}
+
+define i32 @in_commutativity_1_1_0(i32 %x, i32 %y, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_commutativity_1_1_0:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    movl %esi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_commutativity_1_1_0:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    movl %esi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %x, %n1 ; swapped, %x instead of %y
+  ret i32 %r
+}
+
+define i32 @in_commutativity_1_1_1(i32 %x, i32 %y, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_commutativity_1_1_1:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    movl %esi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_commutativity_1_1_1:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    movl %esi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0 ; swapped
+  %r = xor i32 %x, %n1 ; swapped, %x instead of %y
+  ret i32 %r
+}
+
+; ============================================================================ ;
+; Constant Commutativity tests.
+; ============================================================================ ;
+
+define i32 @in_constmask_commutativity_0_1(i32 %x, i32 %y) {
+; CHECK-NOBMI-LABEL: in_constmask_commutativity_0_1:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_constmask_commutativity_0_1:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %y, %n1 ; swapped
+  ret i32 %r
+}
+
+define i32 @in_constmask_commutativity_1_0(i32 %x, i32 %y) {
+; CHECK-NOBMI-LABEL: in_constmask_commutativity_1_0:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    andl $16776960, %esi # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    movl %esi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_constmask_commutativity_1_0:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    andl $16776960, %esi # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    movl %esi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %n1, %x ; %x instead of %y
+  ret i32 %r
+}
+
+define i32 @in_constmask_commutativity_1_1(i32 %x, i32 %y) {
+; CHECK-NOBMI-LABEL: in_constmask_commutativity_1_1:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    andl $16776960, %esi # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    xorl %edi, %esi
+; CHECK-NOBMI-NEXT:    movl %esi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_constmask_commutativity_1_1:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    andl $16776960, %esi # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    xorl %edi, %esi
+; CHECK-BMI-NEXT:    movl %esi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %x, %n1 ; swapped, %x instead of %y
+  ret i32 %r
+}
+
+; ============================================================================ ;
+; Y is an 'and' too.
+; ============================================================================ ;
+
+define i32 @in_complex_y0(i32 %x, i32 %y_hi, i32 %y_low, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_complex_y0:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %ecx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_complex_y0:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %ecx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %y = and i32 %y_hi, %y_low
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_complex_y1(i32 %x, i32 %y_hi, i32 %y_low, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_complex_y1:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %ecx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_complex_y1:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %ecx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %y = and i32 %y_hi, %y_low
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %y, %n1
+  ret i32 %r
+}
+
+define i32 @in_complex_y0_constmask(i32 %x, i32 %y_hi, i32 %y_low) {
+; CHECK-NOBMI-LABEL: in_complex_y0_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_complex_y0_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %y = and i32 %y_hi, %y_low
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_complex_y1_constmask(i32 %x, i32 %y_hi, i32 %y_low) {
+; CHECK-NOBMI-LABEL: in_complex_y1_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_complex_y1_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %y = and i32 %y_hi, %y_low
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %y, %n1
+  ret i32 %r
+}
+
+; ============================================================================ ;
+; M is an 'xor' too.
+; ============================================================================ ;
+
+define i32 @in_complex_m0(i32 %x, i32 %y, i32 %m_a, i32 %m_b) {
+; CHECK-NOBMI-LABEL: in_complex_m0:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %ecx, %edx
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_complex_m0:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %ecx, %edx
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_complex_m1(i32 %x, i32 %y, i32 %m_a, i32 %m_b) {
+; CHECK-NOBMI-LABEL: in_complex_m1:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %ecx, %edx
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_complex_m1:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %ecx, %edx
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+; ============================================================================ ;
+; Both Y and M are complex.
+; ============================================================================ ;
+
+define i32 @in_complex_y0_m0(i32 %x, i32 %y_hi, i32 %y_low, i32 %m_a, i32 %m_b) {
+; CHECK-NOBMI-LABEL: in_complex_y0_m0:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %r8d, %ecx
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %ecx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_complex_y0_m0:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %r8d, %ecx
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %ecx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %y = and i32 %y_hi, %y_low
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_complex_y1_m0(i32 %x, i32 %y_hi, i32 %y_low, i32 %m_a, i32 %m_b) {
+; CHECK-NOBMI-LABEL: in_complex_y1_m0:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %r8d, %ecx
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %ecx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_complex_y1_m0:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %r8d, %ecx
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %ecx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %y = and i32 %y_hi, %y_low
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %y, %n1
+  ret i32 %r
+}
+
+define i32 @in_complex_y0_m1(i32 %x, i32 %y_hi, i32 %y_low, i32 %m_a, i32 %m_b) {
+; CHECK-NOBMI-LABEL: in_complex_y0_m1:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %r8d, %ecx
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %ecx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_complex_y0_m1:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %r8d, %ecx
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %ecx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %y = and i32 %y_hi, %y_low
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_complex_y1_m1(i32 %x, i32 %y_hi, i32 %y_low, i32 %m_a, i32 %m_b) {
+; CHECK-NOBMI-LABEL: in_complex_y1_m1:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %esi
+; CHECK-NOBMI-NEXT:    xorl %r8d, %ecx
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %ecx, %edi
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_complex_y1_m1:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %esi
+; CHECK-BMI-NEXT:    xorl %r8d, %ecx
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %ecx, %edi
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %y = and i32 %y_hi, %y_low
+  %mask = xor i32 %m_a, %m_b
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %mask, %n0
+  %r = xor i32 %y, %n1
+  ret i32 %r
+}
+
+; ============================================================================ ;
+; Negative tests. Should not be folded.
+; ============================================================================ ;
+
+; Multi-use tests.
+
+declare void @use32(i32)
+
+define i32 @in_multiuse_A(i32 %x, i32 %y, i32 %z, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_multiuse_A:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    pushq %rbp
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NOBMI-NEXT:    pushq %rbx
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 24
+; CHECK-NOBMI-NEXT:    pushq %rax
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-NOBMI-NEXT:    .cfi_offset %rbx, -24
+; CHECK-NOBMI-NEXT:    .cfi_offset %rbp, -16
+; CHECK-NOBMI-NEXT:    movl %esi, %ebx
+; CHECK-NOBMI-NEXT:    movl %edi, %ebp
+; CHECK-NOBMI-NEXT:    xorl %esi, %ebp
+; CHECK-NOBMI-NEXT:    andl %ecx, %ebp
+; CHECK-NOBMI-NEXT:    movl %ebp, %edi
+; CHECK-NOBMI-NEXT:    callq use32
+; CHECK-NOBMI-NEXT:    xorl %ebx, %ebp
+; CHECK-NOBMI-NEXT:    movl %ebp, %eax
+; CHECK-NOBMI-NEXT:    addq $8, %rsp
+; CHECK-NOBMI-NEXT:    popq %rbx
+; CHECK-NOBMI-NEXT:    popq %rbp
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_multiuse_A:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    pushq %rbp
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-BMI-NEXT:    pushq %rbx
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 24
+; CHECK-BMI-NEXT:    pushq %rax
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-BMI-NEXT:    .cfi_offset %rbx, -24
+; CHECK-BMI-NEXT:    .cfi_offset %rbp, -16
+; CHECK-BMI-NEXT:    movl %esi, %ebx
+; CHECK-BMI-NEXT:    movl %edi, %ebp
+; CHECK-BMI-NEXT:    xorl %esi, %ebp
+; CHECK-BMI-NEXT:    andl %ecx, %ebp
+; CHECK-BMI-NEXT:    movl %ebp, %edi
+; CHECK-BMI-NEXT:    callq use32
+; CHECK-BMI-NEXT:    xorl %ebx, %ebp
+; CHECK-BMI-NEXT:    movl %ebp, %eax
+; CHECK-BMI-NEXT:    addq $8, %rsp
+; CHECK-BMI-NEXT:    popq %rbx
+; CHECK-BMI-NEXT:    popq %rbp
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  call void @use32(i32 %n1)
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_multiuse_A_constmask(i32 %x, i32 %y, i32 %z) {
+; CHECK-NOBMI-LABEL: in_multiuse_A_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    pushq %rbp
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NOBMI-NEXT:    pushq %rbx
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 24
+; CHECK-NOBMI-NEXT:    pushq %rax
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-NOBMI-NEXT:    .cfi_offset %rbx, -24
+; CHECK-NOBMI-NEXT:    .cfi_offset %rbp, -16
+; CHECK-NOBMI-NEXT:    movl %esi, %ebx
+; CHECK-NOBMI-NEXT:    movl %edi, %ebp
+; CHECK-NOBMI-NEXT:    xorl %esi, %ebp
+; CHECK-NOBMI-NEXT:    andl $16776960, %ebp # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    movl %ebp, %edi
+; CHECK-NOBMI-NEXT:    callq use32
+; CHECK-NOBMI-NEXT:    xorl %ebx, %ebp
+; CHECK-NOBMI-NEXT:    movl %ebp, %eax
+; CHECK-NOBMI-NEXT:    addq $8, %rsp
+; CHECK-NOBMI-NEXT:    popq %rbx
+; CHECK-NOBMI-NEXT:    popq %rbp
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_multiuse_A_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    pushq %rbp
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-BMI-NEXT:    pushq %rbx
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 24
+; CHECK-BMI-NEXT:    pushq %rax
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-BMI-NEXT:    .cfi_offset %rbx, -24
+; CHECK-BMI-NEXT:    .cfi_offset %rbp, -16
+; CHECK-BMI-NEXT:    movl %esi, %ebx
+; CHECK-BMI-NEXT:    movl %edi, %ebp
+; CHECK-BMI-NEXT:    xorl %esi, %ebp
+; CHECK-BMI-NEXT:    andl $16776960, %ebp # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    movl %ebp, %edi
+; CHECK-BMI-NEXT:    callq use32
+; CHECK-BMI-NEXT:    xorl %ebx, %ebp
+; CHECK-BMI-NEXT:    movl %ebp, %eax
+; CHECK-BMI-NEXT:    addq $8, %rsp
+; CHECK-BMI-NEXT:    popq %rbx
+; CHECK-BMI-NEXT:    popq %rbp
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  call void @use32(i32 %n1)
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_multiuse_B(i32 %x, i32 %y, i32 %z, i32 %mask) {
+; CHECK-NOBMI-LABEL: in_multiuse_B:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    pushq %rbp
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NOBMI-NEXT:    pushq %rbx
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 24
+; CHECK-NOBMI-NEXT:    pushq %rax
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-NOBMI-NEXT:    .cfi_offset %rbx, -24
+; CHECK-NOBMI-NEXT:    .cfi_offset %rbp, -16
+; CHECK-NOBMI-NEXT:    movl %ecx, %ebx
+; CHECK-NOBMI-NEXT:    movl %esi, %ebp
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %edi, %ebx
+; CHECK-NOBMI-NEXT:    callq use32
+; CHECK-NOBMI-NEXT:    xorl %ebp, %ebx
+; CHECK-NOBMI-NEXT:    movl %ebx, %eax
+; CHECK-NOBMI-NEXT:    addq $8, %rsp
+; CHECK-NOBMI-NEXT:    popq %rbx
+; CHECK-NOBMI-NEXT:    popq %rbp
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_multiuse_B:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    pushq %rbp
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-BMI-NEXT:    pushq %rbx
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 24
+; CHECK-BMI-NEXT:    pushq %rax
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-BMI-NEXT:    .cfi_offset %rbx, -24
+; CHECK-BMI-NEXT:    .cfi_offset %rbp, -16
+; CHECK-BMI-NEXT:    movl %ecx, %ebx
+; CHECK-BMI-NEXT:    movl %esi, %ebp
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %edi, %ebx
+; CHECK-BMI-NEXT:    callq use32
+; CHECK-BMI-NEXT:    xorl %ebp, %ebx
+; CHECK-BMI-NEXT:    movl %ebx, %eax
+; CHECK-BMI-NEXT:    addq $8, %rsp
+; CHECK-BMI-NEXT:    popq %rbx
+; CHECK-BMI-NEXT:    popq %rbp
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  call void @use32(i32 %n0)
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+define i32 @in_multiuse_B_constmask(i32 %x, i32 %y, i32 %z) {
+; CHECK-NOBMI-LABEL: in_multiuse_B_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    pushq %rbp
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NOBMI-NEXT:    pushq %rbx
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 24
+; CHECK-NOBMI-NEXT:    pushq %rax
+; CHECK-NOBMI-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-NOBMI-NEXT:    .cfi_offset %rbx, -24
+; CHECK-NOBMI-NEXT:    .cfi_offset %rbp, -16
+; CHECK-NOBMI-NEXT:    movl %esi, %ebx
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %ebp
+; CHECK-NOBMI-NEXT:    andl $16776960, %ebp # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    callq use32
+; CHECK-NOBMI-NEXT:    xorl %ebx, %ebp
+; CHECK-NOBMI-NEXT:    movl %ebp, %eax
+; CHECK-NOBMI-NEXT:    addq $8, %rsp
+; CHECK-NOBMI-NEXT:    popq %rbx
+; CHECK-NOBMI-NEXT:    popq %rbp
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: in_multiuse_B_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    pushq %rbp
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-BMI-NEXT:    pushq %rbx
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 24
+; CHECK-BMI-NEXT:    pushq %rax
+; CHECK-BMI-NEXT:    .cfi_def_cfa_offset 32
+; CHECK-BMI-NEXT:    .cfi_offset %rbx, -24
+; CHECK-BMI-NEXT:    .cfi_offset %rbp, -16
+; CHECK-BMI-NEXT:    movl %esi, %ebx
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    movl %edi, %ebp
+; CHECK-BMI-NEXT:    andl $16776960, %ebp # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    callq use32
+; CHECK-BMI-NEXT:    xorl %ebx, %ebp
+; CHECK-BMI-NEXT:    movl %ebp, %eax
+; CHECK-BMI-NEXT:    addq $8, %rsp
+; CHECK-BMI-NEXT:    popq %rbx
+; CHECK-BMI-NEXT:    popq %rbp
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  call void @use32(i32 %n0)
+  %r = xor i32 %n1, %y
+  ret i32 %r
+}
+
+; Various bad variants
+
+define i32 @n0_badmask(i32 %x, i32 %y, i32 %mask, i32 %mask2) {
+; CHECK-NOBMI-LABEL: n0_badmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    notl %ecx
+; CHECK-NOBMI-NEXT:    andl %esi, %ecx
+; CHECK-NOBMI-NEXT:    orl %edi, %ecx
+; CHECK-NOBMI-NEXT:    movl %ecx, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: n0_badmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    andnl %esi, %ecx, %eax
+; CHECK-BMI-NEXT:    orl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i32 %x, %mask
+  %notmask = xor i32 %mask2, -1 ; %mask2 instead of %mask
+  %my = and i32 %y, %notmask
+  %r = or i32 %mx, %my
+  ret i32 %r
+}
+
+define i32 @n0_badxor(i32 %x, i32 %y, i32 %mask) {
+; CHECK-NOBMI-LABEL: n0_badxor:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl %edx, %edi
+; CHECK-NOBMI-NEXT:    xorl $1, %edx
+; CHECK-NOBMI-NEXT:    andl %esi, %edx
+; CHECK-NOBMI-NEXT:    orl %edi, %edx
+; CHECK-NOBMI-NEXT:    movl %edx, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: n0_badxor:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl %edx, %edi
+; CHECK-BMI-NEXT:    xorl $1, %edx
+; CHECK-BMI-NEXT:    andl %esi, %edx
+; CHECK-BMI-NEXT:    orl %edi, %edx
+; CHECK-BMI-NEXT:    movl %edx, %eax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i32 %x, %mask
+  %notmask = xor i32 %mask, 1 ; instead of -1
+  %my = and i32 %y, %notmask
+  %r = or i32 %mx, %my
+  ret i32 %r
+}
+
+define i32 @n0_badconstmask(i32 %x, i32 %y) {
+; CHECK-NOBMI-LABEL: n0_badconstmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    andl $-16776960, %esi # imm = 0xFF000100
+; CHECK-NOBMI-NEXT:    orl %edi, %esi
+; CHECK-NOBMI-NEXT:    movl %esi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: n0_badconstmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    andl $-16776960, %esi # imm = 0xFF000100
+; CHECK-BMI-NEXT:    orl %edi, %esi
+; CHECK-BMI-NEXT:    movl %esi, %eax
+; CHECK-BMI-NEXT:    retq
+  %mx = and i32 %x, 16776960
+  %my = and i32 %y, -16776960 ; instead of -16776961
+  %r = or i32 %mx, %my
+  ret i32 %r
+}
+
+define i32 @n1_thirdvar(i32 %x, i32 %y, i32 %z, i32 %mask) {
+; CHECK-NOBMI-LABEL: n1_thirdvar:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl %ecx, %edi
+; CHECK-NOBMI-NEXT:    xorl %edx, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: n1_thirdvar:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl %ecx, %edi
+; CHECK-BMI-NEXT:    xorl %edx, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, %mask
+  %r = xor i32 %n1, %z ; instead of %y
+  ret i32 %r
+}
+
+define i32 @n1_thirdvar_constmask(i32 %x, i32 %y, i32 %z) {
+; CHECK-NOBMI-LABEL: n1_thirdvar_constmask:
+; CHECK-NOBMI:       # %bb.0:
+; CHECK-NOBMI-NEXT:    xorl %esi, %edi
+; CHECK-NOBMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-NOBMI-NEXT:    xorl %edx, %edi
+; CHECK-NOBMI-NEXT:    movl %edi, %eax
+; CHECK-NOBMI-NEXT:    retq
+;
+; CHECK-BMI-LABEL: n1_thirdvar_constmask:
+; CHECK-BMI:       # %bb.0:
+; CHECK-BMI-NEXT:    xorl %esi, %edi
+; CHECK-BMI-NEXT:    andl $16776960, %edi # imm = 0xFFFF00
+; CHECK-BMI-NEXT:    xorl %edx, %edi
+; CHECK-BMI-NEXT:    movl %edi, %eax
+; CHECK-BMI-NEXT:    retq
+  %n0 = xor i32 %x, %y
+  %n1 = and i32 %n0, 16776960
+  %r = xor i32 %n1, %z ; instead of %y
+  ret i32 %r
+}