On Thu, Jun 13, 2024 at 12:08 PM Barry Song <21cnbao@xxxxxxxxx> wrote:
>
> From: Barry Song <v-songbaohua@xxxxxxxx>
>
> The folio_test_anon(folio)==false case within do_swap_page() has been
> relocated to folio_add_new_anon_rmap(). Additionally, two other callers
> consistently pass anonymous folios.
>
> stack 1:
> remove_migration_pmd
> -> folio_add_anon_rmap_pmd
> -> __folio_add_anon_rmap
>
> stack 2:
> __split_huge_pmd_locked
> -> folio_add_anon_rmap_ptes
> -> __folio_add_anon_rmap
>
> __folio_add_anon_rmap() only needs to handle the cases
> folio_test_anon(folio)==true now.
My team reported a case where swapoff() is calling
folio_add_anon_rmap_pte *not* folio_add_anon_rmap_ptes
with one new anon (!folio_test_anon(folio)).
I will double check all folio_add_anon_rmap_pte() cases.
>
> Suggested-by: David Hildenbrand <david@xxxxxxxxxx>
> Signed-off-by: Barry Song <v-songbaohua@xxxxxxxx>
> ---
> mm/rmap.c | 17 +++--------------
> 1 file changed, 3 insertions(+), 14 deletions(-)
>
> diff --git a/mm/rmap.c b/mm/rmap.c
> index e612d999811a..e84c706c8241 100644
> --- a/mm/rmap.c
> +++ b/mm/rmap.c
> @@ -1299,21 +1299,10 @@ static __always_inline void __folio_add_anon_rmap(struct folio *folio,
>
> nr = __folio_add_rmap(folio, page, nr_pages, level, &nr_pmdmapped);
>
> - if (unlikely(!folio_test_anon(folio))) {
> - VM_WARN_ON_FOLIO(!folio_test_locked(folio), folio);
> - /*
> - * For a PTE-mapped large folio, we only know that the single
> - * PTE is exclusive. Further, __folio_set_anon() might not get
> - * folio->index right when not given the address of the head
> - * page.
> - */
> - VM_WARN_ON_FOLIO(folio_test_large(folio) &&
> - level != RMAP_LEVEL_PMD, folio);
> - __folio_set_anon(folio, vma, address,
> - !!(flags & RMAP_EXCLUSIVE));
> - } else if (likely(!folio_test_ksm(folio))) {
> + VM_WARN_ON_FOLIO(!folio_test_anon(folio), folio);
> +
> + if (likely(!folio_test_ksm(folio)))
> __page_check_anon_rmap(folio, page, vma, address);
> - }
>
> __folio_mod_stat(folio, nr, nr_pmdmapped);
>
> --
> 2.34.1
>
