A320 will
always try to correct the uncommanded roll (in normal law, of course).
There is a caveat, though.
The way she tries to do it is nice, easy & smooth. First, it's not the way to fly the airplane in moderate turbulence, second, the inbuilt lag makes legions of busdrivers believe (wrongly) that she doesn't pick up the dropped wing.
Typical encounter with gust that is too much for ELACs and SECs to handle immediately, makes Average Joe Busdriver slam the stick into wind. Bad news is that startled FBW has just started to pick up the wing and now it adds its automatic command to stick request and then plane rolls rapidly into wind, zooming past wings level. Now stick goes to opposite side and nice airplane-pilot coupling in roll ensues - sometimes with the tipical bus sound of stick clackclackclackclackclackclacking against the roll stops. I've seen the guys hitting the stops in 9G14 and I never had to use more than ha'f-a-stick in 17G29. Mind you, in both cases bank was maintained within 5° of wings level. FCOM says pilot should filter his stick inputs. Way I interpret (and do) it is; when the wing drops, I help ELACs by giving approximately half the deflection, but return stick to neutral slightly before wings level. Maybe even this is excessive, but for me works like a charm.
As for rudder: kicking out the drift is too much for FBW to handle but if you squeeze it out, bank will stay constant with stick in neutral.
On rotation, prior to lift-off, the upwind wing invariably rises, even though the pilot is easing off the usual slight downwind rudder.
This is another of A320 quirks, on take-off the rudder has to be briskly moved to neutral as soon as rotation is initiated. In everything except the gustiest of weather, the airplane settles into wind with zero bank and zero stick input as if by magic. First time I've was briefed on this technique, I thought that my instructor was pulling my leg. It turned out that he wasn't and that the thing that lifts the wing is the rudder held too long and not the wind.