diff --git a/mlir/lib/Dialect/Linalg/Transforms/Loops.cpp b/mlir/lib/Dialect/Linalg/Transforms/Loops.cpp
--- a/mlir/lib/Dialect/Linalg/Transforms/Loops.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/Loops.cpp
@@ -333,23 +333,28 @@
 
 template <typename IndexedValueType>
 void emitScalarImplementation(ArrayRef<Value> allIvs, PoolingMaxOp op) {
-  auto indices = getInputAndOutputIndices(allIvs, op);
+  InputAndOutputIndices indices = getInputAndOutputIndices(allIvs, op);
   // Emit scalar form.
-  Value lhs = std_load(op.output(), indices.outputs);
-  Value rhs = std_load(op.input(), indices.inputs);
+  IndexedValueType output(op.output());
+  IndexedValueType input(op.input());
+  Value lhs = output(indices.outputs);
+  Value rhs = input(indices.inputs);
   using edsc::op::sgt;
   Value maxValue = std_select(sgt(lhs, rhs), lhs, rhs);
-  std_store(maxValue, op.output(), indices.outputs);
+  output(indices.outputs) = maxValue;
 }
+
 template <typename IndexedValueType>
 void emitScalarImplementation(ArrayRef<Value> allIvs, PoolingMinOp op) {
-  auto indices = getInputAndOutputIndices(allIvs, op);
+  InputAndOutputIndices indices = getInputAndOutputIndices(allIvs, op);
   // Emit scalar form.
-  Value lhs = std_load(op.output(), indices.outputs);
-  Value rhs = std_load(op.input(), indices.inputs);
+  IndexedValueType output(op.output());
+  IndexedValueType input(op.input());
+  Value lhs = output(indices.outputs);
+  Value rhs = input(indices.inputs);
   using edsc::op::slt;
   Value minValue = std_select(slt(lhs, rhs), lhs, rhs);
-  std_store(minValue, op.output(), indices.outputs);
+  output(indices.outputs) = minValue;
 }
 template <typename IndexedValueType>
 void emitScalarImplementation(ArrayRef<Value> allIvs, PoolingSumOp op) {
diff --git a/mlir/test/Dialect/Linalg/affine.mlir b/mlir/test/Dialect/Linalg/affine.mlir
--- a/mlir/test/Dialect/Linalg/affine.mlir
+++ b/mlir/test/Dialect/Linalg/affine.mlir
@@ -123,3 +123,27 @@
 //       CHECK:       %[[inc:.*]] = mulf %[[va]], %[[vb]] : f32
 //       CHECK:       %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
 //       CHECK:       affine.store %[[res]], %[[mC]][%[[b]], %[[m]], %[[n]]] : memref<?x?x?xf32>
+
+// CHECK-LABEL: func @pooling_max_min
+func @pooling_max_min(%arg0: memref<?x?xf32>,
+                      %arg1: memref<?x?xi32>,
+                      %arg2: memref<?x?xf32>) {
+  linalg.pooling_max(%arg0, %arg1, %arg2) { strides = [2, 1] }:
+    memref<?x?xf32>, memref<?x?xi32>, memref<?x?xf32>
+  linalg.pooling_min(%arg0, %arg1, %arg2) { strides = [2, 1] }:
+    memref<?x?xf32>, memref<?x?xi32>, memref<?x?xf32>
+  return
+}
+// This is a basic check to make sure the right load/stores are used. loops.mlir
+// checks for the rest.
+// CHECK:      affine.load
+// CHECK-NEXT: affine.load
+// CHECK-NEXT: cmpf
+// CHECK-NEXT: select
+// CHECK-NEXT: affine.store
+// The min pooling body.
+// CHECK:      affine.load
+// CHECK-NEXT: affine.load
+// CHECK-NEXT: cmpf
+// CHECK-NEXT: select
+// CHECK-NEXT: affine.store