Remarks on the Reynolds problem of elastohydrodynamic lubrication