linux-kernel - Re: [v3 PATCH 04/11] mm: vmscan: remove memcg_shrinker_map

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAHbLzkqo=bHcrLBPd68teEAtfLcOsZZ+e3Eds9EfGakhDbW8zA@mail.gmail.com>
Date:   Wed, 13 Jan 2021 15:48:27 -0800
From:   Yang Shi <shy828301@...il.com>
To:     Kirill Tkhai <ktkhai@...tuozzo.com>
Cc:     Roman Gushchin <guro@...com>, Shakeel Butt <shakeelb@...gle.com>,
        Dave Chinner <david@...morbit.com>,
        Johannes Weiner <hannes@...xchg.org>,
        Michal Hocko <mhocko@...e.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Linux MM <linux-mm@...ck.org>,
        Linux FS-devel Mailing List <linux-fsdevel@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [v3 PATCH 04/11] mm: vmscan: remove memcg_shrinker_map_size

On Wed, Jan 6, 2021 at 2:16 AM Kirill Tkhai <ktkhai@...tuozzo.com> wrote:
>
> On 06.01.2021 01:58, Yang Shi wrote:
> > Both memcg_shrinker_map_size and shrinker_nr_max is maintained, but actually the
> > map size can be calculated via shrinker_nr_max, so it seems unnecessary to keep both.
> > Remove memcg_shrinker_map_size since shrinker_nr_max is also used by iterating the
> > bit map.
> >
> > Signed-off-by: Yang Shi <shy828301@...il.com>
> > ---
> >  mm/vmscan.c | 12 ++++--------
> >  1 file changed, 4 insertions(+), 8 deletions(-)
> >
> > diff --git a/mm/vmscan.c b/mm/vmscan.c
> > index ddb9f972f856..8da765a85569 100644
> > --- a/mm/vmscan.c
> > +++ b/mm/vmscan.c
> > @@ -185,8 +185,7 @@ static LIST_HEAD(shrinker_list);
> >  static DECLARE_RWSEM(shrinker_rwsem);
> >
> >  #ifdef CONFIG_MEMCG
> > -
> > -static int memcg_shrinker_map_size;
> > +static int shrinker_nr_max;
> >
> >  static void memcg_free_shrinker_map_rcu(struct rcu_head *head)
> >  {
> > @@ -248,7 +247,7 @@ int memcg_alloc_shrinker_maps(struct mem_cgroup *memcg)
> >               return 0;
> >
> >       down_read(&shrinker_rwsem);
> > -     size = memcg_shrinker_map_size;
> > +     size = DIV_ROUND_UP(shrinker_nr_max, BITS_PER_LONG) * sizeof(unsigned long);
> >       for_each_node(nid) {
> >               map = kvzalloc(sizeof(*map) + size, GFP_KERNEL);
> >               if (!map) {
> > @@ -269,7 +268,7 @@ static int memcg_expand_shrinker_maps(int new_id)
> >       struct mem_cgroup *memcg;
> >
> >       size = DIV_ROUND_UP(new_id + 1, BITS_PER_LONG) * sizeof(unsigned long);
> > -     old_size = memcg_shrinker_map_size;
> > +     old_size = DIV_ROUND_UP(shrinker_nr_max, BITS_PER_LONG) * sizeof(unsigned long);
> >       if (size <= old_size)
> >               return 0;
>
> These bunch of DIV_ROUND_UP() looks too complex. Since now all the shrinker maps allocation
> logic in the only file, can't we simplify this to look better? I mean something like below
> to merge in your patch:
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index b951c289ef3a..27b6371a1656 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -247,7 +247,7 @@ int memcg_alloc_shrinker_maps(struct mem_cgroup *memcg)
>                 return 0;
>
>         down_read(&shrinker_rwsem);
> -       size = DIV_ROUND_UP(shrinker_nr_max, BITS_PER_LONG) * sizeof(unsigned long);
> +       size = shrinker_nr_max / BITS_PER_BYTE;

The type of shrinker_maps->map is "unsigned long *", I think we should
do "(shrinker_nr_max / BITS_PER_LONG + 1) * sizeof(unsigned long)".

And the "/ BITS_PER_BYTE" makes calculating the pointer of nr_deferred
array harder in the following patch since the length of the map array
may be not multiple of "unsigned long". Without the nr_deferred array,
this change seems fine.

>         for_each_node(nid) {
>                 map = kvzalloc(sizeof(*map) + size, GFP_KERNEL);
>                 if (!map) {
> @@ -264,13 +264,11 @@ int memcg_alloc_shrinker_maps(struct mem_cgroup *memcg)
>
>  static int memcg_expand_shrinker_maps(int new_id)
>  {
> -       int size, old_size, ret = 0;
> +       int size, old_size, new_nr_max, ret = 0;
>         struct mem_cgroup *memcg;
>
>         size = DIV_ROUND_UP(new_id + 1, BITS_PER_LONG) * sizeof(unsigned long);
> -       old_size = DIV_ROUND_UP(shrinker_nr_max, BITS_PER_LONG) * sizeof(unsigned long);
> -       if (size <= old_size)
> -               return 0;
> +       new_nr_max = size * BITS_PER_BYTE;
>
>         if (!root_mem_cgroup)
>                 goto out;
> @@ -287,6 +285,9 @@ static int memcg_expand_shrinker_maps(int new_id)
>         } while ((memcg = mem_cgroup_iter(NULL, memcg, NULL)) != NULL);
>
>  out:
> +       if (ret == 0)
> +               shrinker_nr_max = new_nr_max;
> +
>         return ret;
>  }
>
> @@ -334,8 +335,6 @@ static int prealloc_memcg_shrinker(struct shrinker *shrinker)
>                         idr_remove(&shrinker_idr, id);
>                         goto unlock;
>                 }
> -
> -               shrinker_nr_max = id + 1;
>         }
>         shrinker->id = id;
>         ret = 0;
>