I think a lot is mixed up here. First of all we must be clear, whether we discuss the problem of "electromagnetic mass" on a classical or a quantum level. The former is in some sense the more difficult question, because there is no consistent solution on the "point-particle level". The problem of the diverging mass in classical "electron theory" is infamous. It was discovered by Lorentz and Abraham around 1900. It results from the fact that when trying to calculate the influence of the electromagnetic field of an accelerated particle on itself, this leads to an infinite self-energy. Already for a point particle at rest the Coulomb field's energy diverges. On the other hand it's clear that the electron's measured mass always refers to an electron with its own electromagnetic field around it. So when calculating the "self-force" on the accelerated electron after regularizing this self-field somehow (e.g., by making the electron a little extended object with a charge distribution rather than a point-charge-##\delta## distribution) you can subtract the divergent part of the self-force in the point-particle limit, by lumping it into the mass and declaring the total resulting mass as the measured mass of the electron. Unfortunately this doesn't solve the problem completely, because the resulting equation of motion is a third-order differential equation, the Lorentz-Abraham equation, which shows a-causal behavior. Extending these ideas to relativistic dynamics also doesn't help. Then you have the Lorentz-Abraham-Dirac (LAD) equation. The best you can do is to use the first-order perturbation theory in the self-force, which leads to the Landau-Lifshitz approximation of the LAD equation.
In the quantum-field theoretical description these (UV-)divergences presist, i.e., if you start with the "bare electron" interacting with the electromagnetic field and calculate the electron's self-energy you also get a divergent result, but you can, order-by-order in perturbation theory, do the trick of lumping these infinities into the constants of the theory, i.e., the wave-function normalization factors, the mass of the electron/positron, and the electromagnetic coupling constant, i.e., you get order-by-order in perturbation theory well-defined results for the coupled dynamics of the em. field with the charged particles, and this "renormalized" QED is among the most accurate theories ever, predicting some properties to 12 or more digits of accuracy (e.g., the anomalous magnetic moment of the electron, the Lamb shift of hydrogen-energy levels).