linux-kernel - Re: [PATCH v3 3/5] mm: memcg: add per-memcg zswap writeback stat

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: <CAJD7tkbz1d-E-q0MphD=676-ftp-bmOOCFgVxq5mRKgbgGLuYA@mail.gmail.com>
Date:   Wed, 18 Oct 2023 16:24:38 -0700
From:   Yosry Ahmed <yosryahmed@...gle.com>
To:     Nhat Pham <nphamcs@...il.com>
Cc:     akpm@...ux-foundation.org, hannes@...xchg.org,
        cerasuolodomenico@...il.com, sjenning@...hat.com,
        ddstreet@...e.org, vitaly.wool@...sulko.com, mhocko@...nel.org,
        roman.gushchin@...ux.dev, shakeelb@...gle.com,
        muchun.song@...ux.dev, linux-mm@...ck.org, kernel-team@...a.com,
        linux-kernel@...r.kernel.org, cgroups@...r.kernel.org,
        linux-doc@...r.kernel.org, linux-kselftest@...r.kernel.org,
        shuah@...nel.org
Subject: Re: [PATCH v3 3/5] mm: memcg: add per-memcg zswap writeback stat

On Tue, Oct 17, 2023 at 4:21 PM Nhat Pham <nphamcs@...il.com> wrote:
>
> From: Domenico Cerasuolo <cerasuolodomenico@...il.com>
>
> Since zswap now writes back pages from memcg-specific LRUs, we now need a
> new stat to show writebacks count for each memcg.
>
> Suggested-by: Nhat Pham <nphamcs@...il.com>
> Signed-off-by: Domenico Cerasuolo <cerasuolodomenico@...il.com>
> Signed-off-by: Nhat Pham <nphamcs@...il.com>
> ---
>  include/linux/memcontrol.h |  2 ++
>  mm/memcontrol.c            | 15 +++++++++++++++
>  mm/zswap.c                 |  3 +++
>  3 files changed, 20 insertions(+)
>
> diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
> index 3de10fabea0f..7868b1e00bf5 100644
> --- a/include/linux/memcontrol.h
> +++ b/include/linux/memcontrol.h
> @@ -38,6 +38,7 @@ enum memcg_stat_item {
>         MEMCG_KMEM,
>         MEMCG_ZSWAP_B,
>         MEMCG_ZSWAPPED,
> +       MEMCG_ZSWAP_WB,
>         MEMCG_NR_STAT,
>  };
>
> @@ -1884,6 +1885,7 @@ static inline void count_objcg_event(struct obj_cgroup *objcg,
>  bool obj_cgroup_may_zswap(struct obj_cgroup *objcg);
>  void obj_cgroup_charge_zswap(struct obj_cgroup *objcg, size_t size);
>  void obj_cgroup_uncharge_zswap(struct obj_cgroup *objcg, size_t size);
> +void obj_cgroup_report_zswap_wb(struct obj_cgroup *objcg);
>  #else
>  static inline bool obj_cgroup_may_zswap(struct obj_cgroup *objcg)
>  {
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 1bde67b29287..a9118871e5a6 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -1505,6 +1505,7 @@ static const struct memory_stat memory_stats[] = {
>  #if defined(CONFIG_MEMCG_KMEM) && defined(CONFIG_ZSWAP)
>         { "zswap",                      MEMCG_ZSWAP_B                   },
>         { "zswapped",                   MEMCG_ZSWAPPED                  },
> +       { "zswap_wb",                   MEMCG_ZSWAP_WB                  },

zswap_writeback would be more consistent with file_writeback below.

Taking a step back, this is not really a "state". We increment it by 1
every time and never decrement it. Sounds awfully similar to events :)

You can also use count_objcg_event() directly and avoid the need for
obj_cgroup_report_zswap_wb() below.

>  #endif
>         { "file_mapped",                NR_FILE_MAPPED                  },
>         { "file_dirty",                 NR_FILE_DIRTY                   },
> @@ -1541,6 +1542,7 @@ static int memcg_page_state_unit(int item)
>         switch (item) {
>         case MEMCG_PERCPU_B:
>         case MEMCG_ZSWAP_B:
> +       case MEMCG_ZSWAP_WB:
>         case NR_SLAB_RECLAIMABLE_B:
>         case NR_SLAB_UNRECLAIMABLE_B:
>         case WORKINGSET_REFAULT_ANON:
> @@ -7861,6 +7863,19 @@ void obj_cgroup_uncharge_zswap(struct obj_cgroup *objcg, size_t size)
>         rcu_read_unlock();
>  }
>
> +void obj_cgroup_report_zswap_wb(struct obj_cgroup *objcg)
> +{
> +       struct mem_cgroup *memcg;
> +
> +       if (!cgroup_subsys_on_dfl(memory_cgrp_subsys))
> +               return;
> +
> +       rcu_read_lock();
> +       memcg = obj_cgroup_memcg(objcg);
> +       mod_memcg_state(memcg, MEMCG_ZSWAP_WB, 1);
> +       rcu_read_unlock();
> +}
> +
>  static u64 zswap_current_read(struct cgroup_subsys_state *css,
>                               struct cftype *cft)
>  {
> diff --git a/mm/zswap.c b/mm/zswap.c
> index d2989ad11814..15485427e3fa 100644
> --- a/mm/zswap.c
> +++ b/mm/zswap.c
> @@ -704,6 +704,9 @@ static enum lru_status shrink_memcg_cb(struct list_head *item, struct list_lru_o
>         }
>         zswap_written_back_pages++;
>
> +       if (entry->objcg)
> +               obj_cgroup_report_zswap_wb(entry->objcg);
> +
>         /*
>          * Writeback started successfully, the page now belongs to the
>          * swapcache. Drop the entry from zswap - unless invalidate already
> --
> 2.34.1