Good that you found a workaround to the issue. It's still a valid bug because a VT switch shouldn't be required to make the resolution changes take. Also, it is a driver issue in fact, since the underlying logic for implementing the resolution changes are inside the driver; this issue doesn't occur on -intel for instance.
Anyway, guess next step is to forward this bug upstream.
Good that you found a workaround to the issue. It's still a valid bug because a VT switch shouldn't be required to make the resolution changes take. Also, it is a driver issue in fact, since the underlying logic for implementing the resolution changes are inside the driver; this issue doesn't occur on -intel for instance.
Anyway, guess next step is to forward this bug upstream.