Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20210816180909.3603-1-shy828301@gmail.com> <20210818063042.GA2310427@u2004>
 <dd225992-ec89-a911-b318-f4a91c70ed42@redhat.com> <20210818075316.GB501393@hori.linux.bs1.fc.nec.co.jp>
 <c0082d37-dd36-5d83-f759-cf05a9e00b77@redhat.com>
In-Reply-To: <c0082d37-dd36-5d83-f759-cf05a9e00b77@redhat.com>
From:   Yang Shi <shy828301@gmail.com>
Date:   Wed, 18 Aug 2021 10:04:42 -0700
Message-ID: <CAHbLzkq7M5aDOxmgHGkoh95bC6DfZqQgdMSKS_m-REM4WbYEGw@mail.gmail.com>
Subject: Re: [PATCH 1/2] mm: hwpoison: don't drop slab caches for offlining
 non-LRU page
To:     David Hildenbrand <david@redhat.com>
Cc:     =?UTF-8?B?SE9SSUdVQ0hJIE5BT1lBKOWggOWPoyDnm7TkuZ8p?= 
        <naoya.horiguchi@nec.com>,
        Naoya Horiguchi <naoya.horiguchi@linux.dev>,
        "osalvador@suse.de" <osalvador@suse.de>,
        "tdmackey@twitter.com" <tdmackey@twitter.com>,
        "akpm@linux-foundation.org" <akpm@linux-foundation.org>,
        "corbet@lwn.net" <corbet@lwn.net>,
        "linux-mm@kvack.org" <linux-mm@kvack.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Aug 18, 2021 at 12:55 AM David Hildenbrand <david@redhat.com> wrote=
:
>
> On 18.08.21 09:53, HORIGUCHI NAOYA(=E5=A0=80=E5=8F=A3 =E7=9B=B4=E4=B9=9F)=
 wrote:
> > On Wed, Aug 18, 2021 at 09:24:01AM +0200, David Hildenbrand wrote:
> > ...
> >>
> >> Might we want to add a TODO in the code? We have a similar one in
> >> mm/page_isolation.c:set_migratetype_isolate() and it's certainly a rem=
inder
> >> that something of value is missing.
> >
> > Yes, that will be helpful.  The below's what's in my mind, but if someo=
ne
> > has better idea, that's fine.
> >
> > @@ -296,11 +296,9 @@ void shake_page(struct page *p, int access)
> >       }
> >
> >       /*
> > -      * Only call shrink_node_slabs here (which would also shrink
> > -      * other caches) if access is not potentially fatal.
> > +      * TODO: Could shrink slab caches here if a lightweight range-bas=
ed
> > +      * shrinker will be available.
> >        */
> > -     if (access)
> > -             drop_slab_node(page_to_nid(p));
> >   }
> >   EXPORT_SYMBOL_GPL(shake_page);
>
> Just what I had in mind, thanks!

Fine to me, will add this in v2.

>
>
> --
> Thanks,
>
> David / dhildenb
>
>