Masking in Pooling layer from input layer?

Question

I need to compute the GlobalAveragePooling1D in a layer with a mask. This mask is coming from the input to the model as a layer. So, I wonder if this is doing right what it