Index: lib/Target/X86/X86InstrCompiler.td
===================================================================
--- lib/Target/X86/X86InstrCompiler.td
+++ lib/Target/X86/X86InstrCompiler.td
@@ -1360,6 +1360,44 @@
 }
 } // AddedComplexity, SchedRW
 
+//===----------------------------------------------------------------------===//
+// Pattern match SUB as XOR
+//===----------------------------------------------------------------------===//
+
+// An immediate in the LHS of a subtract can't be encoded in the instruction.
+// If there is no possibility of a borrow we can use an XOR instead of a SUB
+// to enable the immediate to be folded.
+// TODO: Move this to a DAG combine?
+
+def sub_is_xor : PatFrag<(ops node:$lhs, node:$rhs), (sub node:$lhs, node:$rhs),[{
+  if (ConstantSDNode *CN = dyn_cast<ConstantSDNode>(N->getOperand(0))) {
+    KnownBits Known;
+    CurDAG->computeKnownBits(N->getOperand(1), Known);
+
+    // If all possible ones in the RHS are set in the LHS then there can't be
+    // a borrow and we can use xor.
+    return (~Known.Zero).isSubsetOf(CN->getAPIntValue());
+  }
+
+  return false;
+}]>;
+
+let AddedComplexity = 5 in {
+def : Pat<(sub_is_xor imm:$src2, GR8:$src1),
+          (XOR8ri GR8:$src1, imm:$src2)>;
+def : Pat<(sub_is_xor i16immSExt8:$src2, GR16:$src1),
+          (XOR16ri8 GR16:$src1, i16immSExt8:$src2)>;
+def : Pat<(sub_is_xor imm:$src2, GR16:$src1),
+          (XOR16ri GR16:$src1, imm:$src2)>;
+def : Pat<(sub_is_xor i32immSExt8:$src2, GR32:$src1),
+          (XOR32ri8 GR32:$src1, i32immSExt8:$src2)>;
+def : Pat<(sub_is_xor imm:$src2, GR32:$src1),
+          (XOR32ri GR32:$src1, imm:$src2)>;
+def : Pat<(sub_is_xor i64immSExt8:$src2, GR64:$src1),
+          (XOR64ri8 GR64:$src1, i64immSExt8:$src2)>;
+def : Pat<(sub_is_xor i64immSExt32:$src2, GR64:$src1),
+          (XOR64ri32 GR64:$src1, i64immSExt32:$src2)>;
+}
 
 //===----------------------------------------------------------------------===//
 // Some peepholes
Index: test/CodeGen/X86/bool-math.ll
===================================================================
--- test/CodeGen/X86/bool-math.ll
+++ test/CodeGen/X86/bool-math.ll
@@ -47,8 +47,8 @@
 ; CHECK-LABEL: add_zext_cmp_mask_same_size_result:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    andb $1, %dil
-; CHECK-NEXT:    movb $27, %al
-; CHECK-NEXT:    subb %dil, %al
+; CHECK-NEXT:    xorb $27, %dil
+; CHECK-NEXT:    movl %edi, %eax
 ; CHECK-NEXT:    retq
   %a = and i8 %x, 1
   %c = icmp eq i8 %a, 0
@@ -61,8 +61,8 @@
 ; CHECK-LABEL: add_zext_cmp_mask_wider_result:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    andl $1, %edi
-; CHECK-NEXT:    movl $27, %eax
-; CHECK-NEXT:    subl %edi, %eax
+; CHECK-NEXT:    xorl $27, %edi
+; CHECK-NEXT:    movl %edi, %eax
 ; CHECK-NEXT:    retq
   %a = and i8 %x, 1
   %c = icmp eq i8 %a, 0
@@ -75,8 +75,8 @@
 ; CHECK-LABEL: add_zext_cmp_mask_narrower_result:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    andl $1, %edi
-; CHECK-NEXT:    movb $43, %al
-; CHECK-NEXT:    subb %dil, %al
+; CHECK-NEXT:    xorb $43, %dil
+; CHECK-NEXT:    movl %edi, %eax
 ; CHECK-NEXT:    retq
   %a = and i32 %x, 1
   %c = icmp eq i32 %a, 0
@@ -129,8 +129,8 @@
 ; CHECK-LABEL: low_bit_select_constants_bigger_true_same_size_result:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    andb $1, %dil
-; CHECK-NEXT:    movb $-29, %al
-; CHECK-NEXT:    subb %dil, %al
+; CHECK-NEXT:    xorb $-29, %dil
+; CHECK-NEXT:    movl %edi, %eax
 ; CHECK-NEXT:    retq
   %a = and i8 %x, 1
   %c = icmp eq i8 %a, 0
@@ -142,8 +142,8 @@
 ; CHECK-LABEL: low_bit_select_constants_bigger_true_wider_result:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    andl $1, %edi
-; CHECK-NEXT:    movl $227, %eax
-; CHECK-NEXT:    subl %edi, %eax
+; CHECK-NEXT:    xorl $227, %edi
+; CHECK-NEXT:    movl %edi, %eax
 ; CHECK-NEXT:    retq
   %a = and i8 %x, 1
   %c = icmp eq i8 %a, 0
@@ -155,8 +155,8 @@
 ; CHECK-LABEL: low_bit_select_constants_bigger_true_narrower_result:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    andl $1, %edi
-; CHECK-NEXT:    movb $41, %al
-; CHECK-NEXT:    subb %dil, %al
+; CHECK-NEXT:    xorb $41, %dil
+; CHECK-NEXT:    movl %edi, %eax
 ; CHECK-NEXT:    retq
   %a = and i16 %x, 1
   %c = icmp eq i16 %a, 0
Index: test/CodeGen/X86/dagcombine-select.ll
===================================================================
--- test/CodeGen/X86/dagcombine-select.ll
+++ test/CodeGen/X86/dagcombine-select.ll
@@ -194,9 +194,9 @@
 ; CHECK-LABEL: shl_constant_sel_constants:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    andb $1, %dil
-; CHECK-NEXT:    movb $3, %cl
-; CHECK-NEXT:    subb %dil, %cl
+; CHECK-NEXT:    xorb $3, %dil
 ; CHECK-NEXT:    movl $1, %eax
+; CHECK-NEXT:    movl %edi, %ecx
 ; CHECK-NEXT:    shll %cl, %eax
 ; CHECK-NEXT:    retq
   %sel = select i1 %cond, i32 2, i32 3
@@ -208,9 +208,9 @@
 ; CHECK-LABEL: lshr_constant_sel_constants:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    andb $1, %dil
-; CHECK-NEXT:    movb $3, %cl
-; CHECK-NEXT:    subb %dil, %cl
+; CHECK-NEXT:    xorb $3, %dil
 ; CHECK-NEXT:    movl $64, %eax
+; CHECK-NEXT:    movl %edi, %ecx
 ; CHECK-NEXT:    shrl %cl, %eax
 ; CHECK-NEXT:    retq
   %sel = select i1 %cond, i32 2, i32 3
@@ -222,9 +222,9 @@
 ; CHECK-LABEL: ashr_constant_sel_constants:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    andb $1, %dil
-; CHECK-NEXT:    movb $3, %cl
-; CHECK-NEXT:    subb %dil, %cl
+; CHECK-NEXT:    xorb $3, %dil
 ; CHECK-NEXT:    movl $128, %eax
+; CHECK-NEXT:    movl %edi, %ecx
 ; CHECK-NEXT:    shrl %cl, %eax
 ; CHECK-NEXT:    retq
   %sel = select i1 %cond, i32 2, i32 3
Index: test/CodeGen/X86/extract-lowbits.ll
===================================================================
--- test/CodeGen/X86/extract-lowbits.ll
+++ test/CodeGen/X86/extract-lowbits.ll
@@ -28,34 +28,6 @@
 ; ---------------------------------------------------------------------------- ;
 
 define i32 @bzhi32_a0(i32 %val, i32 %numlowbits) nounwind {
-; X86-NOBMI-LABEL: bzhi32_a0:
-; X86-NOBMI:       # %bb.0:
-; X86-NOBMI-NEXT:    movb {{[0-9]+}}(%esp), %cl
-; X86-NOBMI-NEXT:    movl $1, %eax
-; X86-NOBMI-NEXT:    shll %cl, %eax
-; X86-NOBMI-NEXT:    decl %eax
-; X86-NOBMI-NEXT:    andl {{[0-9]+}}(%esp), %eax
-; X86-NOBMI-NEXT:    retl
-;
-; X86-BMI1BMI2-LABEL: bzhi32_a0:
-; X86-BMI1BMI2:       # %bb.0:
-; X86-BMI1BMI2-NEXT:    movb {{[0-9]+}}(%esp), %al
-; X86-BMI1BMI2-NEXT:    bzhil %eax, {{[0-9]+}}(%esp), %eax
-; X86-BMI1BMI2-NEXT:    retl
-;
-; X64-NOBMI-LABEL: bzhi32_a0:
-; X64-NOBMI:       # %bb.0:
-; X64-NOBMI-NEXT:    movl $1, %eax
-; X64-NOBMI-NEXT:    movl %esi, %ecx
-; X64-NOBMI-NEXT:    shll %cl, %eax
-; X64-NOBMI-NEXT:    decl %eax
-; X64-NOBMI-NEXT:    andl %edi, %eax
-; X64-NOBMI-NEXT:    retq
-;
-; X64-BMI1BMI2-LABEL: bzhi32_a0:
-; X64-BMI1BMI2:       # %bb.0:
-; X64-BMI1BMI2-NEXT:    bzhil %esi, %edi, %eax
-; X64-BMI1BMI2-NEXT:    retq
   %onebit = shl i32 1, %numlowbits
   %mask = add nsw i32 %onebit, -1
   %masked = and i32 %mask, %val