How to compute manually log-likelihood from model deviance with GLM?

I am doing this exercise of a partial on GLM with R. Please consider this GLM () R output:

1. With only this output, how to get the model' log-likelihood? With R it's easy:

   logLik(model)

And it's the same value as deviance/-2

But I'm not sure with this last formula (deviance/-2). Never in my course, it's written this formula.

In my course, to test global model significativity, 2 hypothesis: H0 : Model 1 = Beta0 H1 : Model 2 = Beta0 + Beta1 . X1

And Likelihood test rate = deviance(Model1) - deviance(Model2)

= -2 ln(likelihood Model1 / likelihood Model2)

follows a Khi2 law with a number of df = (model2 df - model1 df) Can you explain me please? Thanks you!