Extremely large-scale multiple-input multiple-output (XL-MIMO) communication aims to further boost the antenna size significantly than current massive MIMO systems, for which conventional far-field assumption with uniform plane wave (UPW) model may become invalid. This paper studies the modelling and performance analysis for multi-user XL-MIMO communication. With the spherical wavefront phase modelling, and also by taking into account the variations of signal amplitude and projected aperture across array elements, the performance of the three typical beamforming schemes are analyzed, namely the maximal-ratio combining (MRC), zero-forcing (ZF), and minimum mean-square error (MMSE) beamforming. For the special case of two-users, we analytically show that the signal-to-interference-plus-noise ratio (SINR) of all the three beamforming schemes increases as the channels' correlation coefficient decreases. Furthermore, compared to existing UPW model where inter-user interference (IUI) can only be suppressed in angular domain, XL-MIMO enables a new degree-of-freedom (DoF) for IUI suppression by distance separation, even for users along the same direction. Simulation results are provided to validate the modelling and performance analysis of multi-user XL-MIMO communications.