diff --git a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
--- a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -3538,6 +3538,11 @@
     }
   }
 
+  // See if we have a sub(X,mul(div(X,Y),Y)) == rem(X,Y) pattern that could be
+  // merged into a divrem.
+  if (SDValue DivRem = useDivRem(N))
+    return DivRem.getValue(1);
+
   // If the relocation model supports it, consider symbol offsets.
   if (GlobalAddressSDNode *GA = dyn_cast<GlobalAddressSDNode>(N0))
     if (!LegalOperations && TLI.isOffsetFoldingLegal(GA)) {
@@ -4021,15 +4026,39 @@
   if (Node->use_empty())
     return SDValue(); // This is a dead node, leave it alone.
 
-  unsigned Opcode = Node->getOpcode();
-  bool isSigned = (Opcode == ISD::SDIV) || (Opcode == ISD::SREM);
-  unsigned DivRemOpc = isSigned ? ISD::SDIVREM : ISD::UDIVREM;
-
   // DivMod lib calls can still work on non-legal types if using lib-calls.
   EVT VT = Node->getValueType(0);
   if (VT.isVector() || !VT.isInteger())
     return SDValue();
 
+  unsigned Opcode = Node->getOpcode();
+  SDValue Op0 = Node->getOperand(0);
+  SDValue Op1 = Node->getOperand(1);
+
+  // Check for a sub(X,mul(div(X,Y),Y)) == rem(X,Y) pattern.
+  if (Opcode == ISD::SUB) {
+    if (Op1.getOpcode() != ISD::MUL)
+      return SDValue();
+    auto IsDiv = [&](SDValue Mul0, SDValue Mul1) {
+      unsigned Opc = Mul0.getOpcode();
+      if ((Opc != ISD::SDIV && Opc != ISD::UDIV) || Mul0.getOperand(0) != Op0 ||
+          Mul0.getOperand(1) != Mul1)
+        return false;
+      Opcode = Opc == ISD::SDIV ? ISD::SREM : ISD::UREM;
+      Op1 = Mul0.getOperand(1);
+      return true;
+    };
+    if (!IsDiv(Op1.getOperand(0), Op1.getOperand(1)) &&
+        !IsDiv(Op1.getOperand(1), Op1.getOperand(0)))
+      return SDValue();
+  }
+
+  assert((Opcode == ISD::SDIV || Opcode == ISD::SREM || Opcode == ISD::UDIV ||
+          Opcode == ISD::UREM) &&
+         "Unknown div/rem opcode");
+  bool isSigned = (Opcode == ISD::SDIV) || (Opcode == ISD::SREM);
+  unsigned DivRemOpc = isSigned ? ISD::SDIVREM : ISD::UDIVREM;
+
   if (!TLI.isTypeLegal(VT) && !TLI.isOperationCustom(DivRemOpc, VT))
     return SDValue();
 
@@ -4051,8 +4080,6 @@
       return SDValue();
   }
 
-  SDValue Op0 = Node->getOperand(0);
-  SDValue Op1 = Node->getOperand(1);
   SDValue combined;
   for (SDNode::use_iterator UI = Op0.getNode()->use_begin(),
          UE = Op0.getNode()->use_end(); UI != UE; ++UI) {
diff --git a/llvm/test/CodeGen/X86/div-rem-pair-recomposition-signed.ll b/llvm/test/CodeGen/X86/div-rem-pair-recomposition-signed.ll
--- a/llvm/test/CodeGen/X86/div-rem-pair-recomposition-signed.ll
+++ b/llvm/test/CodeGen/X86/div-rem-pair-recomposition-signed.ll
@@ -11,25 +11,20 @@
 define i8 @scalar_i8(i8 %x, i8 %y, i8* %divdst) nounwind {
 ; X86-LABEL: scalar_i8:
 ; X86:       # %bb.0:
+; X86-NEXT:    movsbl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    idivb {{[0-9]+}}(%esp)
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %edx
-; X86-NEXT:    movb {{[0-9]+}}(%esp), %ch
-; X86-NEXT:    movb {{[0-9]+}}(%esp), %cl
-; X86-NEXT:    movsbl %cl, %eax
-; X86-NEXT:    idivb %ch
+; X86-NEXT:    movsbl %ah, %ecx
 ; X86-NEXT:    movb %al, (%edx)
-; X86-NEXT:    mulb %ch
-; X86-NEXT:    subb %al, %cl
 ; X86-NEXT:    movl %ecx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: scalar_i8:
 ; X64:       # %bb.0:
-; X64-NEXT:    movsbl %dil, %ecx
-; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    movsbl %dil, %eax
 ; X64-NEXT:    idivb %sil
+; X64-NEXT:    movsbl %ah, %ecx
 ; X64-NEXT:    movb %al, (%rdx)
-; X64-NEXT:    mulb %sil
-; X64-NEXT:    subb %al, %cl
 ; X64-NEXT:    movl %ecx, %eax
 ; X64-NEXT:    retq
   %div = sdiv i8 %x, %y
@@ -42,34 +37,23 @@
 define i16 @scalar_i16(i16 %x, i16 %y, i16* %divdst) nounwind {
 ; X86-LABEL: scalar_i16:
 ; X86:       # %bb.0:
-; X86-NEXT:    pushl %edi
-; X86-NEXT:    pushl %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %edi
-; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    movzwl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    cwtd
-; X86-NEXT:    idivw %si
-; X86-NEXT:    # kill: def $ax killed $ax def $eax
-; X86-NEXT:    movw %ax, (%edi)
-; X86-NEXT:    imull %eax, %esi
-; X86-NEXT:    subl %esi, %ecx
-; X86-NEXT:    movl %ecx, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    popl %edi
+; X86-NEXT:    idivw {{[0-9]+}}(%esp)
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    movw %ax, (%ecx)
+; X86-NEXT:    movl %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: scalar_i16:
 ; X64:       # %bb.0:
 ; X64-NEXT:    movq %rdx, %rcx
 ; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    # kill: def $ax killed $ax killed $eax
 ; X64-NEXT:    cwtd
 ; X64-NEXT:    idivw %si
-; X64-NEXT:    # kill: def $ax killed $ax def $eax
 ; X64-NEXT:    movw %ax, (%rcx)
-; X64-NEXT:    imull %eax, %esi
-; X64-NEXT:    subl %esi, %edi
-; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %edx, %eax
 ; X64-NEXT:    retq
   %div = sdiv i16 %x, %y
   store i16 %div, i16* %divdst, align 4
@@ -81,20 +65,12 @@
 define i32 @scalar_i32(i32 %x, i32 %y, i32* %divdst) nounwind {
 ; X86-LABEL: scalar_i32:
 ; X86:       # %bb.0:
-; X86-NEXT:    pushl %edi
-; X86-NEXT:    pushl %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %edi
-; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    cltd
-; X86-NEXT:    idivl %edi
-; X86-NEXT:    movl %eax, (%esi)
-; X86-NEXT:    imull %edi, %eax
-; X86-NEXT:    subl %eax, %ecx
-; X86-NEXT:    movl %ecx, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    popl %edi
+; X86-NEXT:    idivl {{[0-9]+}}(%esp)
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    movl %eax, (%ecx)
+; X86-NEXT:    movl %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: scalar_i32:
@@ -104,9 +80,7 @@
 ; X64-NEXT:    cltd
 ; X64-NEXT:    idivl %esi
 ; X64-NEXT:    movl %eax, (%rcx)
-; X64-NEXT:    imull %esi, %eax
-; X64-NEXT:    subl %eax, %edi
-; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %edx, %eax
 ; X64-NEXT:    retq
   %div = sdiv i32 %x, %y
   store i32 %div, i32* %divdst, align 4
@@ -158,9 +132,7 @@
 ; X64-NEXT:    cqto
 ; X64-NEXT:    idivq %rsi
 ; X64-NEXT:    movq %rax, (%rcx)
-; X64-NEXT:    imulq %rsi, %rax
-; X64-NEXT:    subq %rax, %rdi
-; X64-NEXT:    movq %rdi, %rax
+; X64-NEXT:    movq %rdx, %rax
 ; X64-NEXT:    retq
   %div = sdiv i64 %x, %y
   store i64 %div, i64* %divdst, align 4
@@ -880,34 +852,23 @@
 define i32 @scalar_i32_commutative(i32 %x, i32* %ysrc, i32* %divdst) nounwind {
 ; X86-LABEL: scalar_i32_commutative:
 ; X86:       # %bb.0:
-; X86-NEXT:    pushl %edi
-; X86-NEXT:    pushl %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl (%eax), %edi
-; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X86-NEXT:    cltd
-; X86-NEXT:    idivl %edi
-; X86-NEXT:    movl %eax, (%esi)
-; X86-NEXT:    imull %eax, %edi
-; X86-NEXT:    subl %edi, %ecx
-; X86-NEXT:    movl %ecx, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    popl %edi
+; X86-NEXT:    idivl (%ecx)
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    movl %eax, (%ecx)
+; X86-NEXT:    movl %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: scalar_i32_commutative:
 ; X64:       # %bb.0:
 ; X64-NEXT:    movq %rdx, %rcx
-; X64-NEXT:    movl (%rsi), %esi
 ; X64-NEXT:    movl %edi, %eax
 ; X64-NEXT:    cltd
-; X64-NEXT:    idivl %esi
+; X64-NEXT:    idivl (%rsi)
 ; X64-NEXT:    movl %eax, (%rcx)
-; X64-NEXT:    imull %eax, %esi
-; X64-NEXT:    subl %esi, %edi
-; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %edx, %eax
 ; X64-NEXT:    retq
   %y = load i32, i32* %ysrc, align 4
   %div = sdiv i32 %x, %y
@@ -921,24 +882,20 @@
 define i32 @extrause(i32 %x, i32 %y, i32* %divdst, i32* %t1dst) nounwind {
 ; X86-LABEL: extrause:
 ; X86:       # %bb.0:
-; X86-NEXT:    pushl %ebx
 ; X86-NEXT:    pushl %edi
 ; X86-NEXT:    pushl %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %edi
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ebx
-; X86-NEXT:    movl %ecx, %eax
 ; X86-NEXT:    cltd
-; X86-NEXT:    idivl %ebx
+; X86-NEXT:    idivl %ecx
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %edi
 ; X86-NEXT:    movl %eax, (%edi)
-; X86-NEXT:    imull %ebx, %eax
+; X86-NEXT:    imull %ecx, %eax
 ; X86-NEXT:    movl %eax, (%esi)
-; X86-NEXT:    subl %eax, %ecx
-; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    movl %edx, %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    popl %edi
-; X86-NEXT:    popl %ebx
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: extrause:
@@ -950,8 +907,7 @@
 ; X64-NEXT:    movl %eax, (%r8)
 ; X64-NEXT:    imull %esi, %eax
 ; X64-NEXT:    movl %eax, (%rcx)
-; X64-NEXT:    subl %eax, %edi
-; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %edx, %eax
 ; X64-NEXT:    retq
   %div = sdiv i32 %x, %y
   store i32 %div, i32* %divdst, align 4
diff --git a/llvm/test/CodeGen/X86/div-rem-pair-recomposition-unsigned.ll b/llvm/test/CodeGen/X86/div-rem-pair-recomposition-unsigned.ll
--- a/llvm/test/CodeGen/X86/div-rem-pair-recomposition-unsigned.ll
+++ b/llvm/test/CodeGen/X86/div-rem-pair-recomposition-unsigned.ll
@@ -11,25 +11,20 @@
 define i8 @scalar_i8(i8 %x, i8 %y, i8* %divdst) nounwind {
 ; X86-LABEL: scalar_i8:
 ; X86:       # %bb.0:
+; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    divb {{[0-9]+}}(%esp)
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %edx
-; X86-NEXT:    movb {{[0-9]+}}(%esp), %ch
-; X86-NEXT:    movb {{[0-9]+}}(%esp), %cl
-; X86-NEXT:    movzbl %cl, %eax
-; X86-NEXT:    divb %ch
+; X86-NEXT:    movzbl %ah, %ecx
 ; X86-NEXT:    movb %al, (%edx)
-; X86-NEXT:    mulb %ch
-; X86-NEXT:    subb %al, %cl
 ; X86-NEXT:    movl %ecx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: scalar_i8:
 ; X64:       # %bb.0:
-; X64-NEXT:    movzbl %dil, %ecx
-; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    movzbl %dil, %eax
 ; X64-NEXT:    divb %sil
+; X64-NEXT:    movzbl %ah, %ecx
 ; X64-NEXT:    movb %al, (%rdx)
-; X64-NEXT:    mulb %sil
-; X64-NEXT:    subb %al, %cl
 ; X64-NEXT:    movl %ecx, %eax
 ; X64-NEXT:    retq
   %div = udiv i8 %x, %y
@@ -42,34 +37,23 @@
 define i16 @scalar_i16(i16 %x, i16 %y, i16* %divdst) nounwind {
 ; X86-LABEL: scalar_i16:
 ; X86:       # %bb.0:
-; X86-NEXT:    pushl %edi
-; X86-NEXT:    pushl %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %edi
-; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    movzwl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    xorl %edx, %edx
-; X86-NEXT:    divw %si
-; X86-NEXT:    # kill: def $ax killed $ax def $eax
-; X86-NEXT:    movw %ax, (%edi)
-; X86-NEXT:    imull %eax, %esi
-; X86-NEXT:    subl %esi, %ecx
-; X86-NEXT:    movl %ecx, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    popl %edi
+; X86-NEXT:    divw {{[0-9]+}}(%esp)
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    movw %ax, (%ecx)
+; X86-NEXT:    movl %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: scalar_i16:
 ; X64:       # %bb.0:
 ; X64-NEXT:    movq %rdx, %rcx
 ; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    # kill: def $ax killed $ax killed $eax
 ; X64-NEXT:    xorl %edx, %edx
 ; X64-NEXT:    divw %si
-; X64-NEXT:    # kill: def $ax killed $ax def $eax
 ; X64-NEXT:    movw %ax, (%rcx)
-; X64-NEXT:    imull %eax, %esi
-; X64-NEXT:    subl %esi, %edi
-; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %edx, %eax
 ; X64-NEXT:    retq
   %div = udiv i16 %x, %y
   store i16 %div, i16* %divdst, align 4
@@ -81,20 +65,12 @@
 define i32 @scalar_i32(i32 %x, i32 %y, i32* %divdst) nounwind {
 ; X86-LABEL: scalar_i32:
 ; X86:       # %bb.0:
-; X86-NEXT:    pushl %edi
-; X86-NEXT:    pushl %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %edi
-; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    xorl %edx, %edx
-; X86-NEXT:    divl %edi
-; X86-NEXT:    movl %eax, (%esi)
-; X86-NEXT:    imull %edi, %eax
-; X86-NEXT:    subl %eax, %ecx
-; X86-NEXT:    movl %ecx, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    popl %edi
+; X86-NEXT:    divl {{[0-9]+}}(%esp)
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    movl %eax, (%ecx)
+; X86-NEXT:    movl %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: scalar_i32:
@@ -104,9 +80,7 @@
 ; X64-NEXT:    xorl %edx, %edx
 ; X64-NEXT:    divl %esi
 ; X64-NEXT:    movl %eax, (%rcx)
-; X64-NEXT:    imull %esi, %eax
-; X64-NEXT:    subl %eax, %edi
-; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %edx, %eax
 ; X64-NEXT:    retq
   %div = udiv i32 %x, %y
   store i32 %div, i32* %divdst, align 4
@@ -158,9 +132,7 @@
 ; X64-NEXT:    xorl %edx, %edx
 ; X64-NEXT:    divq %rsi
 ; X64-NEXT:    movq %rax, (%rcx)
-; X64-NEXT:    imulq %rsi, %rax
-; X64-NEXT:    subq %rax, %rdi
-; X64-NEXT:    movq %rdi, %rax
+; X64-NEXT:    movq %rdx, %rax
 ; X64-NEXT:    retq
   %div = udiv i64 %x, %y
   store i64 %div, i64* %divdst, align 4
@@ -880,34 +852,23 @@
 define i32 @scalar_i32_commutative(i32 %x, i32* %ysrc, i32* %divdst) nounwind {
 ; X86-LABEL: scalar_i32_commutative:
 ; X86:       # %bb.0:
-; X86-NEXT:    pushl %edi
-; X86-NEXT:    pushl %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl (%eax), %edi
-; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X86-NEXT:    xorl %edx, %edx
-; X86-NEXT:    divl %edi
-; X86-NEXT:    movl %eax, (%esi)
-; X86-NEXT:    imull %eax, %edi
-; X86-NEXT:    subl %edi, %ecx
-; X86-NEXT:    movl %ecx, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    popl %edi
+; X86-NEXT:    divl (%ecx)
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    movl %eax, (%ecx)
+; X86-NEXT:    movl %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: scalar_i32_commutative:
 ; X64:       # %bb.0:
 ; X64-NEXT:    movq %rdx, %rcx
-; X64-NEXT:    movl (%rsi), %esi
 ; X64-NEXT:    movl %edi, %eax
 ; X64-NEXT:    xorl %edx, %edx
-; X64-NEXT:    divl %esi
+; X64-NEXT:    divl (%rsi)
 ; X64-NEXT:    movl %eax, (%rcx)
-; X64-NEXT:    imull %eax, %esi
-; X64-NEXT:    subl %esi, %edi
-; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %edx, %eax
 ; X64-NEXT:    retq
   %y = load i32, i32* %ysrc, align 4
   %div = udiv i32 %x, %y
@@ -921,24 +882,20 @@
 define i32 @extrause(i32 %x, i32 %y, i32* %divdst, i32* %t1dst) nounwind {
 ; X86-LABEL: extrause:
 ; X86:       # %bb.0:
-; X86-NEXT:    pushl %ebx
 ; X86-NEXT:    pushl %edi
 ; X86-NEXT:    pushl %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %edi
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ebx
-; X86-NEXT:    movl %ecx, %eax
 ; X86-NEXT:    xorl %edx, %edx
-; X86-NEXT:    divl %ebx
+; X86-NEXT:    divl %ecx
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %edi
 ; X86-NEXT:    movl %eax, (%edi)
-; X86-NEXT:    imull %ebx, %eax
+; X86-NEXT:    imull %ecx, %eax
 ; X86-NEXT:    movl %eax, (%esi)
-; X86-NEXT:    subl %eax, %ecx
-; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    movl %edx, %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    popl %edi
-; X86-NEXT:    popl %ebx
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: extrause:
@@ -950,8 +907,7 @@
 ; X64-NEXT:    movl %eax, (%r8)
 ; X64-NEXT:    imull %esi, %eax
 ; X64-NEXT:    movl %eax, (%rcx)
-; X64-NEXT:    subl %eax, %edi
-; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %edx, %eax
 ; X64-NEXT:    retq
   %div = udiv i32 %x, %y
   store i32 %div, i32* %divdst, align 4