On the Bayesianity of minimum risk equivariant estimator for location or scale parameters under a general convex and invariant loss function