Index: llvm/trunk/lib/Transforms/Scalar/InferAddressSpaces.cpp
===================================================================
--- llvm/trunk/lib/Transforms/Scalar/InferAddressSpaces.cpp
+++ llvm/trunk/lib/Transforms/Scalar/InferAddressSpaces.cpp
@@ -500,6 +500,7 @@
   }
 
   // Computes the operands of the new constant expression.
+  bool IsNew = false;
   SmallVector<Constant *, 4> NewOperands;
   for (unsigned Index = 0; Index < CE->getNumOperands(); ++Index) {
     Constant *Operand = CE->getOperand(Index);
@@ -509,6 +510,7 @@
     // bitcast, and getelementptr) do not incur cycles in the data flow graph
     // and (2) this function is called on constant expressions in postorder.
     if (Value *NewOperand = ValueWithNewAddrSpace.lookup(Operand)) {
+      IsNew = true;
       NewOperands.push_back(cast<Constant>(NewOperand));
     } else {
       // Otherwise, reuses the old operand.
@@ -516,6 +518,11 @@
     }
   }
 
+  // If !IsNew, we will replace the Value with itself. However, replaced values
+  // are assumed to wrapped in a addrspace cast later so drop it now.
+  if (!IsNew)
+    return nullptr;
+
   if (CE->getOpcode() == Instruction::GetElementPtr) {
     // Needs to specify the source type while constructing a getelementptr
     // constant expression.
Index: llvm/trunk/test/Transforms/InferAddressSpaces/NVPTX/clone_constexpr.ll
===================================================================
--- llvm/trunk/test/Transforms/InferAddressSpaces/NVPTX/clone_constexpr.ll
+++ llvm/trunk/test/Transforms/InferAddressSpaces/NVPTX/clone_constexpr.ll
@@ -0,0 +1,36 @@
+; RUN: opt -S -mtriple=nvptx64-nvidia-cuda -infer-address-spaces %s | FileCheck %s
+
+%struct.S = type { [5 x i32] }
+
+$g1 = comdat any
+
+@g1 = linkonce_odr addrspace(3) global %struct.S zeroinitializer, comdat, align 4
+
+; CHECK-LABEL: @foo(
+; CHECK:  %x0 = tail call i32 @llvm.nvvm.read.ptx.sreg.tid.x() #2
+; CHECK:  %idxprom.i = zext i32 %x0 to i64
+; CHECK:  %arrayidx.i = getelementptr %struct.S, %struct.S* addrspacecast (%struct.S addrspace(3)* @g1 to %struct.S*), i64 0, i32 0, i64 %idxprom.i
+; CHECK:  tail call void @f1(i32* %arrayidx.i, i32 undef) #0
+; CHECK:  %x1 = load i32, i32* getelementptr (%struct.S, %struct.S* addrspacecast (%struct.S addrspace(3)* @g1 to %struct.S*), i64 0, i32 0, i64 0), align 4
+; CHECK:  %L.sroa.0.0.insert.ext.i = zext i32 %x1 to i64
+; CHECK:  tail call void @f2(i64* null, i64 %L.sroa.0.0.insert.ext.i) #0
+; CHECK:  ret void
+define void @foo() local_unnamed_addr #0 {
+entry:
+  %x0 = tail call i32 @llvm.nvvm.read.ptx.sreg.tid.x() #2
+  %idxprom.i = zext i32 %x0 to i64
+  %arrayidx.i = getelementptr %struct.S, %struct.S* addrspacecast (%struct.S addrspace(3)* @g1 to %struct.S*), i64 0, i32 0, i64 %idxprom.i
+  tail call void @f1(i32* %arrayidx.i, i32 undef) #0
+  %x1 = load i32, i32* getelementptr (%struct.S, %struct.S* addrspacecast (%struct.S addrspace(3)* @g1 to %struct.S*), i64 0, i32 0, i64 0), align 4
+  %L.sroa.0.0.insert.ext.i = zext i32 %x1 to i64
+  tail call void @f2(i64* null, i64 %L.sroa.0.0.insert.ext.i) #0
+  ret void
+}
+
+declare void @f1(i32*, i32) local_unnamed_addr #0
+declare void @f2(i64*, i64) local_unnamed_addr #0
+declare i32 @llvm.nvvm.read.ptx.sreg.tid.x() #1
+
+attributes #0 = { convergent nounwind }
+attributes #1 = { nounwind readnone }
+attributes #2 = { nounwind }