Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.35 as permitted sender) client-ip=23.128.96.35;
MIME-Version: 1.0
References: <20230922193639.10158-1-vishal.moola@gmail.com>
 <20230922193639.10158-3-vishal.moola@gmail.com> <CAHbLzkqt8LFFDH2a7+xaQmncwiK=Ynv6vHs9d=TDNjp9_wJ4rw@mail.gmail.com>
In-Reply-To: <CAHbLzkqt8LFFDH2a7+xaQmncwiK=Ynv6vHs9d=TDNjp9_wJ4rw@mail.gmail.com>
From:   Yang Shi <shy828301@gmail.com>
Date:   Fri, 29 Sep 2023 12:07:35 -0700
Message-ID: <CAHbLzkqK+V5EhJ7JgA11VnBPYKD9=HNB0rLJ2_1J-s34BtNeTw@mail.gmail.com>
Subject: Re: [RFC PATCH 2/2] mm/khugepaged: Remove compound_pagelist
To:     "Vishal Moola (Oracle)" <vishal.moola@gmail.com>
Cc:     linux-mm@kvack.org, akpm@linux-foundation.org,
        linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Tue, Sep 26, 2023 at 3:07=E2=80=AFPM Yang Shi <shy828301@gmail.com> wrot=
e:
>
> On Fri, Sep 22, 2023 at 9:33=E2=80=AFPM Vishal Moola (Oracle)
> <vishal.moola@gmail.com> wrote:
> >
> > Currently, khugepaged builds a compound_pagelist while scanning, which
> > is used to properly account for compound pages. We can now account
> > for a compound page as a singular folio instead, so remove this list.
> >
> > Large folios are guaranteed to have consecutive ptes and addresses, so
> > once the first pte of a large folio is found skip over the rest.
>
> The address space may just map a partial folio, for example, in the
> extreme case the HUGE_PMD size range may have HUGE_PMD_NR folios with
> mapping one subpage from each folio per PTE. So assuming the PTE
> mapped folio is mapped consecutively may be wrong.
>
> Please refer to collapse_compound_extreme() in
> tools/testing/selftests/mm/khugepaged.c.
>
> >
> > This helps convert khugepaged to use folios. It removes 3 compound_head
> > calls in __collapse_huge_page_copy_succeeded(), and removes 980 bytes o=
f
> > kernel text.
> >
> > Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
> > ---
> >  mm/khugepaged.c | 76 ++++++++++++-------------------------------------
> >  1 file changed, 18 insertions(+), 58 deletions(-)
> >
> > diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> > index f46a7a7c489f..b6c7d55a8231 100644
> > --- a/mm/khugepaged.c
> > +++ b/mm/khugepaged.c
> > @@ -498,10 +498,9 @@ static void release_pte_page(struct page *page)
> >         release_pte_folio(page_folio(page));
> >  }
> >
> > -static void release_pte_pages(pte_t *pte, pte_t *_pte,
> > -               struct list_head *compound_pagelist)
> > +static void release_pte_folios(pte_t *pte, pte_t *_pte)
> >  {
> > -       struct folio *folio, *tmp;
> > +       struct folio *folio;
> >
> >         while (--_pte >=3D pte) {
> >                 pte_t pteval =3D ptep_get(_pte);
> > @@ -514,12 +513,7 @@ static void release_pte_pages(pte_t *pte, pte_t *_=
pte,
> >                         continue;
> >                 folio =3D pfn_folio(pfn);
> >                 if (folio_test_large(folio))
> > -                       continue;
> > -               release_pte_folio(folio);
> > -       }
> > -
> > -       list_for_each_entry_safe(folio, tmp, compound_pagelist, lru) {
> > -               list_del(&folio->lru);
> > +                       _pte -=3D folio_nr_pages(folio) - 1;
> >                 release_pte_folio(folio);
> >         }
> >  }
> > @@ -538,8 +532,7 @@ static bool is_refcount_suitable(struct page *page)
> >  static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
> >                                         unsigned long address,
> >                                         pte_t *pte,
> > -                                       struct collapse_control *cc,
> > -                                       struct list_head *compound_page=
list)
> > +                                       struct collapse_control *cc)
> >  {
> >         struct folio *folio =3D NULL;
> >         pte_t *_pte;
> > @@ -588,19 +581,6 @@ static int __collapse_huge_page_isolate(struct vm_=
area_struct *vma,
> >                         }
> >                 }
> >
> > -               if (folio_test_large(folio)) {
> > -                       struct folio *f;
> > -
> > -                       /*
> > -                        * Check if we have dealt with the compound pag=
e
> > -                        * already
> > -                        */
> > -                       list_for_each_entry(f, compound_pagelist, lru) =
{
> > -                               if (folio =3D=3D f)
> > -                                       goto next;
> > -                       }
> > -               }
> > -
> >                 /*
> >                  * We can do it before isolate_lru_page because the
> >                  * page can't be freed from under us. NOTE: PG_lock
> > @@ -644,9 +624,6 @@ static int __collapse_huge_page_isolate(struct vm_a=
rea_struct *vma,
> >                 VM_BUG_ON_FOLIO(!folio_test_locked(folio), folio);
> >                 VM_BUG_ON_FOLIO(folio_test_lru(folio), folio);
> >
> > -               if (folio_test_large(folio))
> > -                       list_add_tail(&folio->lru, compound_pagelist);
> > -next:
> >                 /*
> >                  * If collapse was initiated by khugepaged, check that =
there is
> >                  * enough young pte to justify collapsing the page
> > @@ -660,6 +637,10 @@ static int __collapse_huge_page_isolate(struct vm_=
area_struct *vma,
> >                 if (pte_write(pteval))
> >                         writable =3D true;
> >
> > +               if (folio_test_large(folio)) {
> > +                       _pte +=3D folio_nr_pages(folio) - 1;
> > +                       address +=3D folio_size(folio) - PAGE_SIZE;
> > +               }
> >         }
> >
> >         if (unlikely(!writable)) {
> > @@ -673,7 +654,7 @@ static int __collapse_huge_page_isolate(struct vm_a=
rea_struct *vma,
> >                 return result;
> >         }
> >  out:
> > -       release_pte_pages(pte, _pte, compound_pagelist);
> > +       release_pte_folios(pte, _pte);
> >         trace_mm_collapse_huge_page_isolate(&folio->page, none_or_zero,
> >                                             referenced, writable, resul=
t);
> >         return result;
> > @@ -682,11 +663,9 @@ static int __collapse_huge_page_isolate(struct vm_=
area_struct *vma,
> >  static void __collapse_huge_page_copy_succeeded(pte_t *pte,
> >                                                 struct vm_area_struct *=
vma,
> >                                                 unsigned long address,
> > -                                               spinlock_t *ptl,
> > -                                               struct list_head *compo=
und_pagelist)
> > +                                               spinlock_t *ptl)
> >  {
> >         struct page *src_page;
> > -       struct page *tmp;
> >         pte_t *_pte;
> >         pte_t pteval;
> >
> > @@ -706,8 +685,7 @@ static void __collapse_huge_page_copy_succeeded(pte=
_t *pte,
> >                         }
> >                 } else {
> >                         src_page =3D pte_page(pteval);
> > -                       if (!PageCompound(src_page))
> > -                               release_pte_page(src_page);
> > +                       release_pte_page(src_page);

This line is problematic too. It may cause double unlock if I read it
correctly. The loop scans the mapped subpages from the same folio,
release_pte_page() is called for the same folio multiple times.

> >                         /*
> >                          * ptl mostly unnecessary, but preempt has to
> >                          * be disabled to update the per-cpu stats
> > @@ -720,23 +698,12 @@ static void __collapse_huge_page_copy_succeeded(p=
te_t *pte,
> >                         free_page_and_swap_cache(src_page);
> >                 }
> >         }
> > -
> > -       list_for_each_entry_safe(src_page, tmp, compound_pagelist, lru)=
 {
> > -               list_del(&src_page->lru);
> > -               mod_node_page_state(page_pgdat(src_page),
> > -                                   NR_ISOLATED_ANON + page_is_file_lru=
(src_page),
> > -                                   -compound_nr(src_page));
> > -               unlock_page(src_page);
> > -               free_swap_cache(src_page);
> > -               putback_lru_page(src_page);
> > -       }
> >  }
> >
> >  static void __collapse_huge_page_copy_failed(pte_t *pte,
> >                                              pmd_t *pmd,
> >                                              pmd_t orig_pmd,
> > -                                            struct vm_area_struct *vma=
,
> > -                                            struct list_head *compound=
_pagelist)
> > +                                            struct vm_area_struct *vma=
)
> >  {
> >         spinlock_t *pmd_ptl;
> >
> > @@ -753,7 +720,7 @@ static void __collapse_huge_page_copy_failed(pte_t =
*pte,
> >          * Release both raw and compound pages isolated
> >          * in __collapse_huge_page_isolate.
> >          */
> > -       release_pte_pages(pte, pte + HPAGE_PMD_NR, compound_pagelist);
> > +       release_pte_folios(pte, pte + HPAGE_PMD_NR);
> >  }
> >
> >  /*
> > @@ -769,7 +736,6 @@ static void __collapse_huge_page_copy_failed(pte_t =
*pte,
> >   * @vma: the original raw pages' virtual memory area
> >   * @address: starting address to copy
> >   * @ptl: lock on raw pages' PTEs
> > - * @compound_pagelist: list that stores compound pages
> >   */
> >  static int __collapse_huge_page_copy(pte_t *pte,
> >                                      struct page *page,
> > @@ -777,8 +743,7 @@ static int __collapse_huge_page_copy(pte_t *pte,
> >                                      pmd_t orig_pmd,
> >                                      struct vm_area_struct *vma,
> >                                      unsigned long address,
> > -                                    spinlock_t *ptl,
> > -                                    struct list_head *compound_pagelis=
t)
> > +                                    spinlock_t *ptl)
> >  {
> >         struct page *src_page;
> >         pte_t *_pte;
> > @@ -804,11 +769,9 @@ static int __collapse_huge_page_copy(pte_t *pte,
> >         }
> >
> >         if (likely(result =3D=3D SCAN_SUCCEED))
> > -               __collapse_huge_page_copy_succeeded(pte, vma, address, =
ptl,
> > -                                                   compound_pagelist);
> > +               __collapse_huge_page_copy_succeeded(pte, vma, address, =
ptl);
> >         else
> > -               __collapse_huge_page_copy_failed(pte, pmd, orig_pmd, vm=
a,
> > -                                                compound_pagelist);
> > +               __collapse_huge_page_copy_failed(pte, pmd, orig_pmd, vm=
a);
> >
> >         return result;
> >  }
> > @@ -1081,7 +1044,6 @@ static int collapse_huge_page(struct mm_struct *m=
m, unsigned long address,
> >                               int referenced, int unmapped,
> >                               struct collapse_control *cc)
> >  {
> > -       LIST_HEAD(compound_pagelist);
> >         pmd_t *pmd, _pmd;
> >         pte_t *pte;
> >         pgtable_t pgtable;
> > @@ -1168,8 +1130,7 @@ static int collapse_huge_page(struct mm_struct *m=
m, unsigned long address,
> >
> >         pte =3D pte_offset_map_lock(mm, &_pmd, address, &pte_ptl);
> >         if (pte) {
> > -               result =3D __collapse_huge_page_isolate(vma, address, p=
te, cc,
> > -                                                     &compound_pagelis=
t);
> > +               result =3D __collapse_huge_page_isolate(vma, address, p=
te, cc);
> >                 spin_unlock(pte_ptl);
> >         } else {
> >                 result =3D SCAN_PMD_NULL;
> > @@ -1198,8 +1159,7 @@ static int collapse_huge_page(struct mm_struct *m=
m, unsigned long address,
> >         anon_vma_unlock_write(vma->anon_vma);
> >
> >         result =3D __collapse_huge_page_copy(pte, hpage, pmd, _pmd,
> > -                                          vma, address, pte_ptl,
> > -                                          &compound_pagelist);
> > +                                          vma, address, pte_ptl);
> >         pte_unmap(pte);
> >         if (unlikely(result !=3D SCAN_SUCCEED))
> >                 goto out_up_write;
> > --
> > 2.40.1
> >