On Thu, Oct 30, 2014 at 02:17:37PM +0200, Kirill A. Shutemov wrote: > On Wed, Oct 29, 2014 at 05:35:17PM +0100, Andrea Arcangeli wrote: > > diff --git a/mm/gup.c b/mm/gup.c > > index a8521f1..01534ff 100644 > > --- a/mm/gup.c > > +++ b/mm/gup.c > > @@ -591,9 +591,9 @@ static __always_inline long __get_user_pages_locked(struct task_struct *tsk, > > int write, int force, > > struct page **pages, > > struct vm_area_struct **vmas, > > - int *locked, bool notify_drop) > > + int *locked, bool notify_drop, > > + unsigned int flags) > > Argument list getting too long. Should we consider packing them into a > struct? It's __always_inline, so it's certainly not a runtime concern. The whole point of using __always_inline is to optimize away certain branches at build time. If this about cleaning it up and not for changing the runtime (which I think couldn't get any better because of the __always_inline), we should at least make certain gcc can still see through the structure offsets to delete the same code blocks at build time if possible, before doing the change. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>