On Tue, Jul 03, 2018 at 01:50:00PM -0700, Andrew Morton wrote:
> On Tue, 03 Jul 2018 18:09:26 +0300 Kirill Tkhai <ktkhai@xxxxxxxxxxxxx> wrote:
>
> > Imagine a big node with many cpus, memory cgroups and containers.
> > Let we have 200 containers, every container has 10 mounts,
> > and 10 cgroups. All container tasks don't touch foreign
> > containers mounts. If there is intensive pages write,
> > and global reclaim happens, a writing task has to iterate
> > over all memcgs to shrink slab, before it's able to go
> > to shrink_page_list().
> >
> > Iteration over all the memcg slabs is very expensive:
> > the task has to visit 200 * 10 = 2000 shrinkers
> > for every memcg, and since there are 2000 memcgs,
> > the total calls are 2000 * 2000 = 4000000.
> >
> > So, the shrinker makes 4 million do_shrink_slab() calls
> > just to try to isolate SWAP_CLUSTER_MAX pages in one
> > of the actively writing memcg via shrink_page_list().
> > I've observed a node spending almost 100% in kernel,
> > making useless iteration over already shrinked slab.
> >
> > This patch adds bitmap of memcg-aware shrinkers to memcg.
> > The size of the bitmap depends on bitmap_nr_ids, and during
> > memcg life it's maintained to be enough to fit bitmap_nr_ids
> > shrinkers. Every bit in the map is related to corresponding
> > shrinker id.
> >
> > Next patches will maintain set bit only for really charged
> > memcg. This will allow shrink_slab() to increase its
> > performance in significant way. See the last patch for
> > the numbers.
> >
> > ...
> >
> > --- a/mm/vmscan.c
> > +++ b/mm/vmscan.c
> > @@ -182,6 +182,11 @@ static int prealloc_memcg_shrinker(struct shrinker *shrinker)
> > if (id < 0)
> > goto unlock;
> >
> > + if (memcg_expand_shrinker_maps(id)) {
> > + idr_remove(&shrinker_idr, id);
> > + goto unlock;
> > + }
> > +
> > if (id >= shrinker_nr_max)
> > shrinker_nr_max = id + 1;
> > shrinker->id = id;
>
> This function ends up being a rather sad little thing.
>
> : static int prealloc_memcg_shrinker(struct shrinker *shrinker)
> : {
> : int id, ret = -ENOMEM;
> :
> : down_write(&shrinker_rwsem);
> : id = idr_alloc(&shrinker_idr, shrinker, 0, 0, GFP_KERNEL);
> : if (id < 0)
> : goto unlock;
> :
> : if (memcg_expand_shrinker_maps(id)) {
> : idr_remove(&shrinker_idr, id);
> : goto unlock;
> : }
> :
> : if (id >= shrinker_nr_max)
> : shrinker_nr_max = id + 1;
> : shrinker->id = id;
> : ret = 0;
> : unlock:
> : up_write(&shrinker_rwsem);
> : return ret;
> : }
>
> - there's no need to call memcg_expand_shrinker_maps() unless id >=
> shrinker_nr_max so why not move the code and avoid calling
> memcg_expand_shrinker_maps() in most cases.
memcg_expand_shrinker_maps will return immediately if per memcg shrinker
maps can accommodate the new id. Since prealloc_memcg_shrinker is
definitely not a hot path, I don't see any penalty in calling this
function on each prealloc_memcg_shrinker invocation.
>
> - why aren't we decreasing shrinker_nr_max in
> unregister_memcg_shrinker()? That's easy to do, avoids pointless
> work in shrink_slab_memcg() and avoids memory waste in future
> prealloc_memcg_shrinker() calls.
We can shrink the maps, but IMHO it isn't worth the complexity it would
introduce, because in my experience if a workload used N mount points
(containers, whatever) at some point of its lifetime, it is likely to
use the same amount in the future.
>
> It should be possible to find the highest ID in an IDR tree with a
> straightforward descent of the underlying radix tree, but I doubt if
> that has been wired up. Otherwise a simple loop in
> unregister_memcg_shrinker() would be needed.
>
>
