[mlir][sparse] Move some member functions from SparseTensorDescriptorImpl to MutSparseTensorDescriptor.

This is to prepare for implementing AOS optimization.

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D141002

mlir/lib/Dialect/SparseTensor/Transforms/SparseTensorCodegen.cpp

@@ -967,9 +967,9 @@ public:
   LogicalResult
   matchAndRewrite(NumberOfEntriesOp op, OpAdaptor adaptor,
                   ConversionPatternRewriter &rewriter) const override {
-    // Query memSizes for the actually stored values size.
-    auto desc = getDescriptorFromTensorTuple(adaptor.getTensor());
-    rewriter.replaceOp(op, desc.getValMemSize(rewriter, op.getLoc()));
+    // Query memSizes for the actually stored values.
+    rewriter.replaceOp(
+        op, genValMemSize(rewriter, op.getLoc(), adaptor.getTensor()));
     return success();
   }
 };