Gotchas related to using SparkSession attached to a Dataframe / Dataset

Question

Are there any disadvantages to using the SparkSession attached to a dataframe / dataset - for example dataframe.sparkSession - as opposed to using the Spark Ses