# GARCH(1,1) regression in Eviews

I'm having a problem in doing a GARCH(1,1) regression. I'm trying to regress gold prices serie on stock returns series as in the following equation in eviews: $$r = c(1) + c(2) \cdot s + c(3)\cdot d10\cdot s +c(4)\cdot d05\cdot s + c(5)\cdot d01\cdot s ,$$ where $r$ is gold prices, $s$ stands for stock prices and $d$ represents dummy variables to capture the 10% 5% and 1% high volatility.

My questions are :

1. Should any of the two series be stationary before I do the regression?
2. Should I write my equation in "the mean equation" or in "the variance equation" or both ?

I'd really appreciate if someone could tell me the stages I should go through to solve this problem, I'd be deeply indepted. Note that I'm a beginner in this GARCH modelling, the more I read about it the more I get confused.