Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.35 as permitted sender) client-ip=23.128.96.35;
From:   "Huang, Ying" <ying.huang@intel.com>
To:     Chris Li <chrisl@kernel.org>
Cc:     Yosry Ahmed <yosryahmed@google.com>,
        Zhongkun He <hezhongkun.hzk@bytedance.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Nhat Pham <nphamcs@gmail.com>,
        Seth Jennings <sjenning@redhat.com>,
        Dan Streetman <ddstreet@ieee.org>,
        Vitaly Wool <vitaly.wool@konsulko.com>,
        linux-mm <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] mm:zswap: fix zswap entry reclamation failure in two
 scenarios
In-Reply-To: <CAF8kJuPs=O7dKQYWAyM5N1UPyCkEmaCb9REHmNvAi7o4tPbMDg@mail.gmail.com>
        (Chris Li's message of "Sun, 19 Nov 2023 21:31:30 -0800")
References: <20231113130601.3350915-1-hezhongkun.hzk@bytedance.com>
        <CAJD7tkY8SwROmNEaBAhkS4OKj33g-6fHsKFeYKW3afT+yAbvxA@mail.gmail.com>
        <CAF8kJuPonfuOtipdifXwBny2H7cy6m6BL8mWFVXzfb9JSdYq3Q@mail.gmail.com>
        <CAJD7tkYMiJiXTTgAN34TP8QTr-ViAuEFddYes=ac+1ErenjCZw@mail.gmail.com>
        <CAF8kJuOC30feLGs0bNHOxMjSZ3uqF1y7eUdJ4p-w-myP8c1cFg@mail.gmail.com>
        <8734x1cdtr.fsf@yhuang6-desk2.ccr.corp.intel.com>
        <CAF8kJuPs=O7dKQYWAyM5N1UPyCkEmaCb9REHmNvAi7o4tPbMDg@mail.gmail.com>
Date:   Mon, 20 Nov 2023 13:39:35 +0800
Message-ID: <87v89xasq0.fsf@yhuang6-desk2.ccr.corp.intel.com>
User-Agent: Gnus/5.13 (Gnus v5.13)
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Chris Li <chrisl@kernel.org> writes:

> On Sun, Nov 19, 2023 at 7:20=E2=80=AFPM Huang, Ying <ying.huang@intel.com=
> wrote:
>>
>> Chris Li <chriscli@google.com> writes:
>>
>> > On Thu, Nov 16, 2023 at 12:19=E2=80=AFPM Yosry Ahmed <yosryahmed@googl=
e.com> wrote:
>> > That will cancel the batching effect on the swap slot free, making the
>> > common case for  swapping  faults take longer to complete, righ?
>> > If I recall correctly, the uncharge is the expensive part of the swap
>> > slot free operation.
>> > I just want to figure out what we are trading off against. This is not
>> > one side wins all situations.
>>
>> Per my understanding, we don't batch memcg uncharging in
>> swap_entry_free() now.  Although it's possible and may improve
>> performance.
>
> swap_entry_free() does not do batching, it is at the caller level.
>
> I just checked. The batching is done in free_swap_slot() is still
> using swap slot cache and batching.
> It uses swapcache_free_entries() to batch free the swap_slots. That is
> where the uncharge happens per my understanding.

Per my understanding, memcg uncharging happens in

swapcache_free_entries()
  swap_entry_free()
    mem_cgroup_uncharge_swap()

The swap entries are uncharged one-by-one, not

acquire lock in memcg
uncharge all entries
release lock in memcg

--
Best Regards,
Huang, Ying