This issue was identified while investigating BF-35256.

A race condition occurs between the __conn_set_timestamp and __conn_reconfigure functions when using the same default session of the connection and accessing the same scratch buffer slot, leading to a heap-use-after-free error when one thread frees the slot while the other tries to access it.