We comment on the macroscopic model for surface plasmons of H.-Y. Deng [New J. Phys. 21 (2019) 043055; arXiv:1712.06101] and a claim, based on energy conversion from charges to the electric field, that surface plasmons on metallic surfaces may become unstable [J. Phys.: Cond. Matt. 29 (2017) 455002; arXiv:1606.06239, 1701.01060]. The discussion revolves around the formulation of charge conservation in the bulk and the surface of a metal. We elaborate in particular on the role of a finite electric current normal to the surface. Using a scheme of Cercignani and Lampis and of Zaremba, we point out that the model chosen by Deng for the non-specular scattering of electrons needs to be amended to prevent the disappearance of charges at the surface. Different models and approaches in the literature on surface plasmons are reviewed: the interfacial excess field approach of Bedeaux and Vlieger which contains Deng's macroscopic model, the assumption of specular reflection of Ritchie and Marusak, a hydrodynamic model with a composite charge density (partially localized at the surface), the local dielectric model, and a macroscopic method with (anti)symmetric fictitious stimuli (used, e.g., by García-Moliner and Flores). This puts Deng's results into perspective and illustrates problems with his approach.