SR is a classical theory, so it can be (and probably should be) considered without dealing with photons, but rather with pulses of light. With any of the classical definitions of the phase of a wave, the phase of the wave will propagate at the speed of light. . I.e. if you pick out a point on a plane wave that has a known value of the E-field (which is the part of the wave that I assume you're measuring the phase of) then whatever the value the E-field and it's phase has will propagate with the speed of light. Thus the phase will be constant along the worldline of the light pulse . If we consider a light pulse propagating in the +x direction, then if the phase is zero at t=0 and x=0, the phase will be zero at t=1sec, x=c*1 sec, or t=2 sec and x=c*2sec, in general at t=a and x=c*a the phase will be the same as it was at t=0 and x=0.