Could you explain to me why you'll give this guarantee?
I am fully aware, that in this specific setup this maybe can help to bind the routing, in other setups it could make the service unavailable over WLAN.
Just roll the change back, and everything is as before, no harm no problem. Just a wrong assumption and dead end testing.
But I like the approach from @admon:
We could test a turn off SMBv1 and after confirming, make a qualified feature request to GL.iNet.
I am on the road now and could not read the smb config to give further instructions right now. Depending on the traffic, maybe in 5 hours.