Received-SPF: pass (google.com: domain of linux-kernel+bounces-79881-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Precedence: bulk
MIME-Version: 1.0
References: <CAGsJ_4w5JF8zjsM=vEiArUpnvN9LjAdnUn98mEUJSAhg4Y9XjQ@mail.gmail.com>
 <20240224205453.47096-1-sj@kernel.org>
In-Reply-To: <20240224205453.47096-1-sj@kernel.org>
From: Barry Song <21cnbao@gmail.com>
Date: Sun, 25 Feb 2024 10:54:24 +1300
Message-ID: <CAGsJ_4x5JcpsdobM88ZnZA4B9nHBEdAfpNcdyqxa5fMDqEMrHQ@mail.gmail.com>
Subject: Re: [PATCH RFC] mm: madvise: pageout: ignore references rather than
 clearing young
To: SeongJae Park <sj@kernel.org>
Cc: Minchan Kim <minchan@kernel.org>, akpm@linux-foundation.org, damon@lists.linux.dev, 
	linux-mm@kvack.org, linux-kernel@vger.kernel.org, mhocko@suse.com, 
	hannes@cmpxchg.org, Barry Song <v-songbaohua@oppo.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Sun, Feb 25, 2024 at 9:54=E2=80=AFAM SeongJae Park <sj@kernel.org> wrote=
:
>
> On Sun, 25 Feb 2024 04:01:40 +0800 Barry Song <21cnbao@gmail.com> wrote:
>
> > On Sun, Feb 25, 2024 at 3:07=E2=80=AFAM SeongJae Park <sj@kernel.org> w=
rote:
> > >
> > > Hi Barry,
> > >
> > > On Sat, 24 Feb 2024 12:37:59 +0800 Barry Song <21cnbao@gmail.com> wro=
te:
> > >
> > > [...]
> > > >
> > > > BTW=EF=BC=8C
> > > > Hi SeongJae,
> > > > I am not quite sure if damon also needs this, so I have kept damon =
as is by
> > > > setting ignore_references =3D false.  MADV_PAGEOUT is an explicit h=
int users
> > > > don't want the memory to be reclaimed, I don't know if it is true f=
or damon as
> > > > well. If you have some comments, please chime in.
> > >
> > > Thank you for calling my name :)
> > >
> > > For DAMON's usecase, the document simply says the behavior would be s=
ame to
> > > MADV_PAGEOUT, so if we conclude to change MADV_PAGEOUT, I think same =
change
> > > should be made for DAMON's usecase, or update DAMON document.
> >
> > Hi SeongJae,
> >
> > I don't find similar clearing pte young in damon_pa_pageout(), so i
> > guess damon's
> > behaviour is actually different with MADV_PAGEOUT which has  pte-cleari=
ng. damon
> > is probably the best-effort but MADV_PAGEOUT isn't .
> >
> > static unsigned long damon_pa_pageout(struct damon_region *r, struct da=
mos *s)
> > {
> >         unsigned long addr, applied;
> >         LIST_HEAD(folio_list);
> >
> >         for (addr =3D r->ar.start; addr < r->ar.end; addr +=3D PAGE_SIZ=
E) {
> >                      struct folio *folio =3D damon_get_folio(PHYS_PFN(a=
ddr));
> >                      ....
> >
> >                      if (damos_pa_filter_out(s, folio))
> >                                 goto put_folio;
> >
> >                     folio_clear_referenced(folio);
> >                     folio_test_clear_young(folio);
> >                     if (!folio_isolate_lru(folio))
> >                          goto put_folio;
> >                     if (folio_test_unevictable(folio))
> >                           folio_putback_lru(folio);
> >                     else
> >                           list_add(&folio->lru, &folio_list);
> > put_folio:
> >                folio_put(folio);
> >          }
> >          applied =3D reclaim_pages(&folio_list);
> >          cond_resched();
> >          return applied * PAGE_SIZE;
> > }
> >
> > am i missing something?
>
> Thank you for checking this again.  You're right.
>
> Technically speaking, DAMON's usage of MADV_PAGEOUT is in vaddr.c.  paddr=
c is
> using not MADV_PAGEOUT but reclaim_pages().  Usage of reclaim_pages() fro=
m
> paddr is different from that of MADV_PAGEOUT since paddr doesn't clear PT=
E.  I
> was confused from the difference between vaddr and paddr.  I actually wan=
ted to
> document the difference but haven't had a time for that yet.  Thank you f=
or
> letting me remind this.

Hi SeongJae,

thanks! I bravely had a go at fixing the damon's doc[1]. as it seems
the fix is anyway needed
no matter if we have my patch to optimize MADV_PAGEOUT.

[1] https://lore.kernel.org/linux-mm/20240224215023.5271-1-21cnbao@gmail.co=
m/

>
> So, your change on MADV_PAGEOUT will make an effect to vaddr, and I think=
 it's
> ok.  Your change on reclaim_pages() could make an effect to paddr, depend=
ing on
> the additional parameter's value.  I now think it would better to make no
> effect here.  That is, let's keep the change for paddr.c in your patch as=
 is.

thanks! it seems everything is quite clear now.

>
>
> Thanks,
> SJ
>
> >
> > >
> > >
> > > Thanks,
> > > SJ
> > >
> >

Thanks
Barry