Received-SPF: pass (google.com: domain of linux-kernel+bounces-159484-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Precedence: bulk
MIME-Version: 1.0
References: <20240425211136.486184-1-zi.yan@sent.com> <CAGsJ_4wa0LskQkoZf9r5bG5+wEkyfCYveMBSTbuDe0=t1QetTg@mail.gmail.com>
 <6C31DF81-94FB-4D09-A3B8-0CED2AD8EDDB@nvidia.com> <CAGsJ_4xzb8RrEuPEbnvR4GbDWuoGCYL4FsC3TObOifAZ4CHGOA@mail.gmail.com>
 <730660D2-E1BA-4A2E-B99C-2F160F9D9A9B@nvidia.com> <CAGsJ_4yNSKdft-G=4X2Dr5HJ-axh-8SohGW7nkUfTYrsNFr9_Q@mail.gmail.com>
 <FAE058F4-7D2A-4760-9D24-A7D931062200@nvidia.com>
In-Reply-To: <FAE058F4-7D2A-4760-9D24-A7D931062200@nvidia.com>
From: Barry Song <21cnbao@gmail.com>
Date: Fri, 26 Apr 2024 11:44:35 +0800
Message-ID: <CAGsJ_4xNARZg29Ay6BWFZ99PnH1k0M2SdPV=jo6gW-uReLfk6Q@mail.gmail.com>
Subject: Re: [PATCH v4] mm/rmap: do not add fully unmapped large folio to
 deferred split list
To: Zi Yan <ziy@nvidia.com>
Cc: Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org, 
	"Matthew Wilcox (Oracle)" <willy@infradead.org>, Yang Shi <shy828301@gmail.com>, 
	Ryan Roberts <ryan.roberts@arm.com>, David Hildenbrand <david@redhat.com>, 
	Lance Yang <ioworker0@gmail.com>, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Fri, Apr 26, 2024 at 11:37=E2=80=AFAM Zi Yan <ziy@nvidia.com> wrote:
>
> On 25 Apr 2024, at 23:28, Barry Song wrote:
>
> > On Fri, Apr 26, 2024 at 10:50=E2=80=AFAM Zi Yan <ziy@nvidia.com> wrote:
> >>
> >> On 25 Apr 2024, at 22:23, Barry Song wrote:
> >>
> >>> On Fri, Apr 26, 2024 at 9:55=E2=80=AFAM Zi Yan <ziy@nvidia.com> wrote=
:
> >>>>
> >>>> On 25 Apr 2024, at 21:45, Barry Song wrote:
> >>>>
> >>>>> On Fri, Apr 26, 2024 at 5:11=E2=80=AFAM Zi Yan <zi.yan@sent.com> wr=
ote:
> >>>>>>
> >>>>>> From: Zi Yan <ziy@nvidia.com>
> >>>>>>
> >>>>>> In __folio_remove_rmap(), a large folio is added to deferred split=
 list
> >>>>>> if any page in a folio loses its final mapping. But it is possible=
 that
> >>>>>> the folio is fully unmapped and adding it to deferred split list i=
s
> >>>>>> unnecessary.
> >>>>>>
> >>>>>> For PMD-mapped THPs, that was not really an issue, because removin=
g the
> >>>>>> last PMD mapping in the absence of PTE mappings would not have add=
ed the
> >>>>>> folio to the deferred split queue.
> >>>>>>
> >>>>>> However, for PTE-mapped THPs, which are now more prominent due to =
mTHP,
> >>>>>> they are always added to the deferred split queue. One side effect
> >>>>>> is that the THP_DEFERRED_SPLIT_PAGE stat for a PTE-mapped folio ca=
n be
> >>>>>> unintentionally increased, making it look like there are many part=
ially
> >>>>>> mapped folios -- although the whole folio is fully unmapped stepwi=
se.
> >>>>>>
> >>>>>> Core-mm now tries batch-unmapping consecutive PTEs of PTE-mapped T=
HPs
> >>>>>> where possible starting from commit b06dc281aa99 ("mm/rmap: introd=
uce
> >>>>>> folio_remove_rmap_[pte|ptes|pmd]()"). When it happens, a whole PTE=
-mapped
> >>>>>> folio is unmapped in one go and can avoid being added to deferred =
split
> >>>>>> list, reducing the THP_DEFERRED_SPLIT_PAGE noise. But there will s=
till be
> >>>>>> noise when we cannot batch-unmap a complete PTE-mapped folio in on=
e go
> >>>>>> -- or where this type of batching is not implemented yet, e.g., mi=
gration.
> >>>>>>
> >>>>>> To avoid the unnecessary addition, folio->_nr_pages_mapped is chec=
ked
> >>>>>> to tell if the whole folio is unmapped. If the folio is already on
> >>>>>> deferred split list, it will be skipped, too.
> >>>>>>
> >>>>>> Note: commit 98046944a159 ("mm: huge_memory: add the missing
> >>>>>> folio_test_pmd_mappable() for THP split statistics") tried to excl=
ude
> >>>>>> mTHP deferred split stats from THP_DEFERRED_SPLIT_PAGE, but it doe=
s not
> >>>>>> fix the above issue. A fully unmapped PTE-mapped order-9 THP was s=
till
> >>>>>> added to deferred split list and counted as THP_DEFERRED_SPLIT_PAG=
E,
> >>>>>> since nr is 512 (non zero), level is RMAP_LEVEL_PTE, and inside
> >>>>>> deferred_split_folio() the order-9 folio is folio_test_pmd_mappabl=
e().
> >>>>>>
> >>>>>> Signed-off-by: Zi Yan <ziy@nvidia.com>
> >>>>>> Reviewed-by: Yang Shi <shy828301@gmail.com>
> >>>>>> ---
> >>>>>>  mm/rmap.c | 8 +++++---
> >>>>>>  1 file changed, 5 insertions(+), 3 deletions(-)
> >>>>>>
> >>>>>> diff --git a/mm/rmap.c b/mm/rmap.c
> >>>>>> index a7913a454028..220ad8a83589 100644
> >>>>>> --- a/mm/rmap.c
> >>>>>> +++ b/mm/rmap.c
> >>>>>> @@ -1553,9 +1553,11 @@ static __always_inline void __folio_remove_=
rmap(struct folio *folio,
> >>>>>>                  * page of the folio is unmapped and at least one =
page
> >>>>>>                  * is still mapped.
> >>>>>>                  */
> >>>>>> -               if (folio_test_large(folio) && folio_test_anon(fol=
io))
> >>>>>> -                       if (level =3D=3D RMAP_LEVEL_PTE || nr < nr=
_pmdmapped)
> >>>>>> -                               deferred_split_folio(folio);
> >>>>>> +               if (folio_test_large(folio) && folio_test_anon(fol=
io) &&
> >>>>>> +                   list_empty(&folio->_deferred_list) &&
> >>>>>> +                   ((level =3D=3D RMAP_LEVEL_PTE && atomic_read(m=
apped)) ||
> >>>>>> +                    (level =3D=3D RMAP_LEVEL_PMD && nr < nr_pmdma=
pped)))
> >>>>>> +                       deferred_split_folio(folio);
> >>>>>
> >>>>> Hi Zi Yan,
> >>>>> in case a mTHP is mapped by two processed (forked but not CoW yet),=
 if we
> >>>>> unmap the whole folio by pte level in one process only, are we stil=
l adding this
> >>>>> folio into deferred list?
> >>>>
> >>>> No. Because the mTHP is still fully mapped by the other process. In =
terms of code,
> >>>> nr will be 0 in that case and this if condition is skipped. nr is on=
ly increased
> >>>> from 0 when one of the subpages in the mTHP has no mapping, namely p=
age->_mapcount
> >>>> becomes negative and last is true in the case RMAP_LEVEL_PTE.
> >>>
> >>> Ok. i see, so "last" won't be true?
> >>>
> >>> case RMAP_LEVEL_PTE:
> >>> do {
> >>> last =3D atomic_add_negative(-1, &page->_mapcount);
> >>>    if (last && folio_test_large(folio)) {
> >>>        last =3D atomic_dec_return_relaxed(mapped);
> >>>        last =3D (last < ENTIRELY_MAPPED);
> >>> }
> >>>
> >>> if (last)
> >>>      nr++;
> >>> } while (page++, --nr_pages > 0);
> >>> break;
> >>
> >> Right, because for every subpage its corresponding
> >> last =3D atomic_add_negative(-1, &page->_mapcount); is not true after =
the unmapping.2
> >
> > if a mTHP is mapped only by one process, and we unmap it entirely, we w=
ill
> > get nr > 0, then we are executing adding it into deferred_list? so it s=
eems
> > atomic_read(mapped) is preventing this case from adding deferred_list?
>
> Yes, that is what this patch is doing. When a mTHP is mapped by one proce=
ss
> and later unmapped fully, there is no need to add it to deferred_list.
> The mTHP will be freed right afterwards.

thanks. I understand. i feel fixing up nr earlier can make the code
more readable.
 case RMAP_LEVEL_PTE:
..
+ if (!atomic_read(mapped))
+     nr =3D 0;
break;

as I have been struggling for a long time to understand the code, especiall=
y
the one with many conditions in the =E2=80=9Cif=E2=80=9D    :-)

+               if (folio_test_large(folio) && folio_test_anon(folio) &&
+                   list_empty(&folio->_deferred_list) &&
+                   ((level =3D=3D RMAP_LEVEL_PTE && atomic_read(mapped)) |=
|
+                    (level =3D=3D RMAP_LEVEL_PMD && nr < nr_pmdmapped)))
+                       deferred_split_folio(folio);
        }

>
> >
> > I wonder if  it is possible to fixup nr to 0 from the first place?
> > for example
> > /* we are doing an entire unmapping */
> > if (page=3D=3D&folio->page && nr_pages =3D=3D  folio_nr_pages(folio))
> > ...
> >
> >>
> >>
> >> --
> >> Best Regards,
> >> Yan, Zi
>
>
> --
> Best Regards,
> Yan, Zi