diff --git a/lib/CodeGen/TargetInfo.cpp b/lib/CodeGen/TargetInfo.cpp
index 0552965..23d3047 100644
--- a/lib/CodeGen/TargetInfo.cpp
+++ b/lib/CodeGen/TargetInfo.cpp
@@ -4044,7 +4044,15 @@ ABIArgInfo AArch64ABIInfo::classifyReturnType(QualType RetTy) const {
   // Aggregates <= 16 bytes are returned directly in registers or on the stack.
   uint64_t Size = getContext().getTypeSize(RetTy);
   if (Size <= 128) {
+    unsigned Alignment = getContext().getTypeAlign(RetTy);
     Size = 64 * ((Size + 63) / 64); // round up to multiple of 8 bytes
+
+    // We use a pair of i64 for 16-byte aggregate with 8-byte alignment.
+    // For aggregates with 16-byte alignment, we use i128.
+    if (Alignment < 128 && Size == 128) {
+      llvm::Type *BaseTy = llvm::Type::getInt64Ty(getVMContext());
+      return ABIArgInfo::getDirect(llvm::ArrayType::get(BaseTy, Size / 64));
+    }
     return ABIArgInfo::getDirect(llvm::IntegerType::get(getVMContext(), Size));
   }
 
diff --git a/test/CodeGen/arm64-arguments.c b/test/CodeGen/arm64-arguments.c
index c276350..4486bb4 100644
--- a/test/CodeGen/arm64-arguments.c
+++ b/test/CodeGen/arm64-arguments.c
@@ -92,7 +92,7 @@ struct s21 f21(void) {}
 // CHECK: define i64 @f22()
 // CHECK: define i64 @f23()
 // CHECK: define i64 @f24()
-// CHECK: define i128 @f25()
+// CHECK: define [2 x i64] @f25()
 // CHECK: define { float, float } @f26()
 // CHECK: define { double, double } @f27()
 _Complex char       f22(void) {}