On Mon, Oct 2, 2017 at 12:56 AM, Michal Hocko <mho...@kernel.org> wrote: > balance_dirty_pages_ratelimited doesn't really need mapping as well. All > it needs is the inode and we already have it in callers. So would it > make sense to refactor a bit further and make its argument an inode?
My only concern is that, balance_dirty_pages_ratelimited() is an exported function so changing its signature could potentially break some drivers?
