Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <1684737363-31554-1-git-send-email-zhaoyang.huang@unisoc.com> <ZG+/EW9NzGsAcGfF@casper.infradead.org>
In-Reply-To: <ZG+/EW9NzGsAcGfF@casper.infradead.org>
From:   Zhaoyang Huang <huangzhaoyang@gmail.com>
Date:   Fri, 26 May 2023 10:30:20 +0800
Message-ID: <CAGWkznGz6w-56wye-0XG=bZMPLY21g33mcAHuWqYvpaHQJUJEw@mail.gmail.com>
Subject: Re: [PATCHv4] mm: skip CMA pages when they are not available
To:     Matthew Wilcox <willy@infradead.org>
Cc:     "zhaoyang.huang" <zhaoyang.huang@unisoc.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Minchan Kim <minchan@kernel.org>,
        Joonsoo Kim <iamjoonsoo.kim@lge.com>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org, ke.wang@unisoc.com
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Fri, May 26, 2023 at 4:03=E2=80=AFAM Matthew Wilcox <willy@infradead.org=
> wrote:
>
> On Mon, May 22, 2023 at 02:36:03PM +0800, zhaoyang.huang wrote:
> > +#ifdef CONFIG_CMA
> > +/*
> > + * It is waste of effort to scan and reclaim CMA pages if it is not av=
ailable
> > + * for current allocation context
> > + */
> > +static bool skip_cma(struct folio *folio, struct scan_control *sc)
> > +{
> > +     if (!current_is_kswapd() &&
> > +                     gfp_migratetype(sc->gfp_mask) !=3D MIGRATE_MOVABL=
E &&
> > +                     get_pageblock_migratetype(&folio->page) =3D=3D MI=
GRATE_CMA)
> > +             return true;
> > +     return false;
> > +}
> > +#else
> > +static bool skip_cma(struct folio *folio, struct scan_control *sc)
> > +{
> > +     return false;
> > +}
> > +#endif
> > +
> >  /*
> >   * Isolating page from the lruvec to fill in @dst list by nr_to_scan t=
imes.
> >   *
> > @@ -2239,7 +2259,8 @@ static unsigned long isolate_lru_folios(unsigned =
long nr_to_scan,
> >               nr_pages =3D folio_nr_pages(folio);
> >               total_scan +=3D nr_pages;
> >
> > -             if (folio_zonenum(folio) > sc->reclaim_idx) {
> > +             if (folio_zonenum(folio) > sc->reclaim_idx ||
> > +                             skip_cma(folio, sc)) {
> >                       nr_skipped[folio_zonenum(folio)] +=3D nr_pages;
> >                       move_to =3D &folios_skipped;
> >                       goto move;
>
> I have no idea if what this patch is trying to accomplish is correct,
> but I no longer object to how it is doing it.
IMO, this is necessary as there could be such weird scenario, that is
an GFP_KERNEL allocation might get 32 MIGRATE_CMA pages via
direct_reclaim which lead to a low PSI_MEM/vmpressure value but return
a NULL pointer