An integral approach to plasma-wall interaction (PWI) modelling for DEMO is presented, which is part of the EUROfusion Theory and Advanced Simulation Coordination activities that were established to advance the understanding and predictive capabilities for the modelling of existing and future fusion devices using a modern advanced computing approach. In view of the DEMO design, the aim of PWI modelling activities is to assess safety-relevant information regarding the erosion of plasma-facing components (PFCs), including its impact on plasma contamination, dust production, fuel inventory, and material response to transient events. This is achieved using a set of powerful and validated computer codes that deal with particular PWI aspects and interact with each other by means of relevant data exchange. Steady state erosion of tungsten PFC and subsequent transport and re-deposition of eroded material are simulated with the ERO2.0 code using a DEMO plasma background produced by dedicated SOLPS-ITER simulations. Dust transport simulations in steady state plasma also rely on the respective SOLPS-ITER solutions and are performed with the MIGRAINe code. In order to improve simulations of tungsten erosion in the divertor of DEMO, relevant high density sheath models are being developed based on particle-in-cell (PIC) simulations with the state-of-the-art BIT code family. PIC codes of the SPICE code family, in turn, provide relevant information on multi-emissive sheath physics, such as semi-empirical scaling laws for field-assisted thermionic emission. These scaling laws are essential for simulations of material melting under transient heat loads that are performed with the recently developed MEMENTO code, the successor of MEMOS-U. Fuel retention simulations assess tritium retention in tungsten and structural materials, as well as fuel permeation to the coolant, accounting for neutron damage. Simulations for divertor monoblocks of different sizes are performed using the FESTIM code, while for the first wall the TESSIM code is applied. Respective code-code dependencies and interactions, as well as modelling results achieved to date are discussed in this contribution.