In the Galí’s book, he linearizes all the equations but he does not use log-deviations
from steady state. In Dynare, can I use the equations in the log-linear format? Is the steady-state zero anyway?

The loglinearization in the book appears to be log form, but it can be easily transformed into log deviation forms only by subtracting the log steady states at both sides!

What exactly do you mean? You would need to compute the steady state first (as always) and then take the logs. Note that log linearizations are more straightforward than Uhlig’s method. See e.g. Log Linearisation Question

Professor, i have a question on this topic. What if the steady state value doesn’t exist when i add some constants as in the Gali’s book? I used basically the same code as in your homepage(for sticky wage chapter) and just added some constant terms like rho in the interest rate rule and natural interest rate and others on the natural wage and so on as described in the book.