Panel Data Unbalanced - Xtreg - Two Way Fix Effect

Hello,

Stata question –

I need help with the following:
I have a panel data with ~7,000 observation include ~15 years (1998-2013) and ~1,500 firms. Not balanced.
In order to regress I have to take in account the hierarchy structure of the data – the repeated measures of firms and years.
I think the best will be a two way fix effect that will capture both fix effect of years and firms in one regression (other?).
For panel data I did: xtset firm year
Variables are: y x1 x2 x3

What will be the best way to run a model?
Many thanks,

Doron