Re: [PATCH V2] mm/swap: take folio refcount after testing the LRU flag

[Yu Zhao <yuzhao@google.com>]

On Mon, Aug 12, 2024 at 5:22 AM <yangge1116@126.com> wrote:
>
> From: yangge <yangge1116@126.com>
>
> Whoever passes a folio to __folio_batch_add_and_move() must hold
> a reference, otherwise something else would already be messed up.
> If the folio is referenced, it will not be freed elsewhere, so we
> can safely clear the folio's lru flag. As discussed with David
> in [1], we should take the reference after testing the LRU flag,
> not before.
>
> Link: https://lore.kernel.org/lkml/d41865b4-d6fa-49ba-890a-921eefad27dd@redhat.com/ [1]
> Signed-off-by: yangge <yangge1116@126.com>
> Acked-by: David Hildenbrand <david@redhat.com>
> ---
>  mm/swap.c | 9 ++++-----
>  1 file changed, 4 insertions(+), 5 deletions(-)
>
> V2:
>    Add sanity check suggested by David
>
> diff --git a/mm/swap.c b/mm/swap.c
> index 67a2467..c048659 100644
> --- a/mm/swap.c
> +++ b/mm/swap.c
> @@ -226,12 +226,11 @@ static void __folio_batch_add_and_move(struct folio_batch __percpu *fbatch,
>  {
>         unsigned long flags;
>
> -       folio_get(folio);
> -
> -       if (on_lru && !folio_test_clear_lru(folio)) {
> -               folio_put(folio);
> +       if (on_lru && !folio_test_clear_lru(folio))
>                 return;
> -       }
> +
> +       VM_WARN_ON_ONCE(!folio_ref_count(folio));
> +       folio_get(folio);

No need to check folio_ref_count() here, because folio_get() already
does it with a better check folio_ref_zero_or_close_to_overflow().