On Mon, Sep 22, 2014 at 04:07:34PM +0200, Michal Hocko wrote: > On Thu 18-09-14 19:50:20, Vladimir Davydov wrote: > > The only reason why this function lives in memcontrol.c is that it > > depends on memcg_caches_array_size. However, we can pass the new array > > size immediately to it instead of new_id+1 so that it will be free of > > any memcontrol.c dependencies. > > > > So let's move this function to slab_common.c and make it static. > > Why?
Jumping from memcontrol.c to slab_common.c and then back to memcontrol.c while updating per-memcg caches looks ugly IMO. We can do the update on the slab's side. > besides that the patch does more code reshuffling which should be > documented. I have got lost a bit to be honest. It just makes it sane :-) Currently we walk over all slab caches each time new kmemcg is created even if memcg_limited_groups_array_size doesn't grow and we've actually nothing to do. So it moves cache id allocation stuff to a separate function (memcg_alloc_cache_id) and places the check there so that memcg_update_all_caches is only called when it's really necessary. I'm sorry if it confuses you. I thought the patch isn't big and rather easy to understand :-/ Next time will split better. Thanks, Vladimir -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
