Received-SPF: pass (google.com: domain of linux-kernel+bounces-43834-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240129054551.57728-1-ioworker0@gmail.com> <CAHbLzkqrTPZD_Fz_a5LYa49Tum_Za=J9dX_SRyTr=kVZA26RAg@mail.gmail.com>
 <CAAa6QmSSKFjzdG62MZ5bSrhP19AjPEr339WXaWY=OCLsGWVE8w@mail.gmail.com>
In-Reply-To: <CAAa6QmSSKFjzdG62MZ5bSrhP19AjPEr339WXaWY=OCLsGWVE8w@mail.gmail.com>
From: Lance Yang <ioworker0@gmail.com>
Date: Tue, 30 Jan 2024 10:37:26 +0800
Message-ID: <CAK1f24mFmqTmBcx2YqQ=TwD0xhA1DhLwPhONp18yzqKdC--K=w@mail.gmail.com>
Subject: Re: [PATCH 1/1] mm/khugepaged: bypassing unnecessary scans with
 MMF_DISABLE_THP check
To: "Zach O'Keefe" <zokeefe@google.com>
Cc: Yang Shi <shy828301@gmail.com>, akpm@linux-foundation.org, mhocko@suse.com, 
	david@redhat.com, songmuchun@bytedance.com, peterx@redhat.com, 
	minchan@kernel.org, linux-mm@kvack.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hey Zach,

Thanks for taking time to review!

On Tue, Jan 30, 2024 at 3:04=E2=80=AFAM Zach O'Keefe <zokeefe@google.com> w=
rote:
[...]
> IIUC, there really isn't any correctness race. Claim is just that we

Yes, there is indeed no correctness race.

> can avoid a number of per-vma checks. AFAICT, any task w/
> MMF_DISABLE_THP set will always have each and every vma checked
> (albeit, with a very inexpensive ->vm_mm->flags check)
[...]

IMO, for any task with MMF_DISABLE_THP set, the check
for each VMA can be skipped to avoid redundant operations,
(with a very inexpensive ->mm->flags check)
especially in scenarios with a large address space.

BR,
Lance

On Tue, Jan 30, 2024 at 3:04=E2=80=AFAM Zach O'Keefe <zokeefe@google.com> w=
rote:
>
> On Mon, Jan 29, 2024 at 10:53=E2=80=AFAM Yang Shi <shy828301@gmail.com> w=
rote:
> >
> > On Sun, Jan 28, 2024 at 9:46=E2=80=AFPM Lance Yang <ioworker0@gmail.com=
> wrote:
> > >
> > > khugepaged scans the entire address space in the
> > > background for each given mm, looking for
> > > opportunities to merge sequences of basic pages
> > > into huge pages. However, when an mm is inserted
> > > to the mm_slots list, and the MMF_DISABLE_THP flag
> > > is set later, this scanning process becomes
> > > unnecessary for that mm and can be skipped to avoid
> > > redundant operations, especially in scenarios with
> > > a large address space.
> > >
> > > This commit introduces a check before each scanning
> > > process to test the MMF_DISABLE_THP flag for the
> > > given mm; if the flag is set, the scanning process
> > > is bypassed, thereby improving the efficiency of
> > > khugepaged.
> > >
> > > Signed-off-by: Lance Yang <ioworker0@gmail.com>
> > > ---
> > >  mm/khugepaged.c | 18 ++++++++++++------
> > >  1 file changed, 12 insertions(+), 6 deletions(-)
> > >
> > > diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> > > index 2b219acb528e..d6a700834edc 100644
> > > --- a/mm/khugepaged.c
> > > +++ b/mm/khugepaged.c
> > > @@ -410,6 +410,12 @@ static inline int hpage_collapse_test_exit(struc=
t mm_struct *mm)
> > >         return atomic_read(&mm->mm_users) =3D=3D 0;
> > >  }
> > >
> > > +static inline int hpage_collapse_test_exit_or_disable(struct mm_stru=
ct *mm)
> > > +{
> > > +       return hpage_collapse_test_exit(mm) ||
> > > +              test_bit(MMF_DISABLE_THP, &mm->flags);
> > > +}
> > > +
> > >  void __khugepaged_enter(struct mm_struct *mm)
> > >  {
> > >         struct khugepaged_mm_slot *mm_slot;
> > > @@ -1422,7 +1428,7 @@ static void collect_mm_slot(struct khugepaged_m=
m_slot *mm_slot)
> > >
> > >         lockdep_assert_held(&khugepaged_mm_lock);
> > >
> > > -       if (hpage_collapse_test_exit(mm)) {
> > > +       if (hpage_collapse_test_exit_or_disable(mm)) {
> > >                 /* free mm_slot */
> > >                 hash_del(&slot->hash);
> > >                 list_del(&slot->mm_node);
> > > @@ -2360,7 +2366,7 @@ static unsigned int khugepaged_scan_mm_slot(uns=
igned int pages, int *result,
> > >                 goto breakouterloop_mmap_lock;
> > >
> > >         progress++;
> > > -       if (unlikely(hpage_collapse_test_exit(mm)))
> > > +       if (unlikely(hpage_collapse_test_exit_or_disable(mm)))
> > >                 goto breakouterloop;
> > >
> > >         vma_iter_init(&vmi, mm, khugepaged_scan.address);
> > > @@ -2368,7 +2374,7 @@ static unsigned int khugepaged_scan_mm_slot(uns=
igned int pages, int *result,
> > >                 unsigned long hstart, hend;
> > >
> > >                 cond_resched();
> > > -               if (unlikely(hpage_collapse_test_exit(mm))) {
> > > +               if (unlikely(hpage_collapse_test_exit_or_disable(mm))=
) {
> >
> > The later thp_vma_allowable_order() does check whether MMF_DISABLE_THP
> > is set or not. And the hugepage_vma_revalidate() after re-acquiring
> > mmap_lock does the same check too. The checking in khugepaged should
> > be already serialized with prctl, which takes mmap_lock in write.
>
> IIUC, there really isn't any correctness race. Claim is just that we
> can avoid a number of per-vma checks. AFAICT, any task w/
> MMF_DISABLE_THP set will always have each and every vma checked
> (albeit, with a very inexpensive ->vm_mm->flags check)
>
> Thanks,
> Zach
>
> > >                         progress++;
> > >                         break;
> > >                 }
> > > @@ -2390,7 +2396,7 @@ static unsigned int khugepaged_scan_mm_slot(uns=
igned int pages, int *result,
> > >                         bool mmap_locked =3D true;
> > >
> > >                         cond_resched();
> > > -                       if (unlikely(hpage_collapse_test_exit(mm)))
> > > +                       if (unlikely(hpage_collapse_test_exit_or_disa=
ble(mm)))
> > >                                 goto breakouterloop;
> > >
> > >                         VM_BUG_ON(khugepaged_scan.address < hstart ||
> > > @@ -2408,7 +2414,7 @@ static unsigned int khugepaged_scan_mm_slot(uns=
igned int pages, int *result,
> > >                                 fput(file);
> > >                                 if (*result =3D=3D SCAN_PTE_MAPPED_HU=
GEPAGE) {
> > >                                         mmap_read_lock(mm);
> > > -                                       if (hpage_collapse_test_exit(=
mm))
> > > +                                       if (hpage_collapse_test_exit_=
or_disable(mm))
> > >                                                 goto breakouterloop;
> > >                                         *result =3D collapse_pte_mapp=
ed_thp(mm,
> > >                                                 khugepaged_scan.addre=
ss, false);
> > > @@ -2450,7 +2456,7 @@ static unsigned int khugepaged_scan_mm_slot(uns=
igned int pages, int *result,
> > >          * Release the current mm_slot if this mm is about to die, or
> > >          * if we scanned all vmas of this mm.
> > >          */
> > > -       if (hpage_collapse_test_exit(mm) || !vma) {
> > > +       if (hpage_collapse_test_exit_or_disable(mm) || !vma) {
> > >                 /*
> > >                  * Make sure that if mm_users is reaching zero while
> > >                  * khugepaged runs here, khugepaged_exit will find
> > > --
> > > 2.33.1
> > >