Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:7 as permitted sender) client-ip=2620:137:e000::3:7;
MIME-Version: 1.0
References: <20231119194740.94101-1-ryncsn@gmail.com> <20231119194740.94101-20-ryncsn@gmail.com>
In-Reply-To: <20231119194740.94101-20-ryncsn@gmail.com>
From:   Chris Li <chrisl@kernel.org>
Date:   Sun, 19 Nov 2023 23:04:02 -0800
Message-ID: <CAF8kJuMJMBjn2VGq8aUzv_t4+k0zPW8WvJaraYe-8GRbX0q1fg@mail.gmail.com>
Subject: Re: [PATCH 19/24] shmem, swap: refactor error check on OOM or race
To:     Kairui Song <kasong@tencent.com>
Cc:     linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>,
        "Huang, Ying" <ying.huang@intel.com>,
        David Hildenbrand <david@redhat.com>,
        Hugh Dickins <hughd@google.com>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Matthew Wilcox <willy@infradead.org>,
        Michal Hocko <mhocko@suse.com>, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hi Kairui,

On Sun, Nov 19, 2023 at 11:49=E2=80=AFAM Kairui Song <ryncsn@gmail.com> wro=
te:
>
> From: Kairui Song <kasong@tencent.com>
>
> It should always check if a swap entry is already swaped in on error,
> fix potential false error issue.
>
> Signed-off-by: Kairui Song <kasong@tencent.com>
> ---
>  mm/shmem.c | 10 ++++------
>  1 file changed, 4 insertions(+), 6 deletions(-)
>
> diff --git a/mm/shmem.c b/mm/shmem.c
> index 81d129aa66d1..6154b5b68b8f 100644
> --- a/mm/shmem.c
> +++ b/mm/shmem.c
> @@ -1857,13 +1857,11 @@ static int shmem_swapin_folio(struct inode *inode=
, pgoff_t index,
>         page =3D swapin_page_non_fault(swap, gfp, mpol, ilx, fault_mm, &r=
esult);
>         mpol_cond_put(mpol);
>
> -       if (PTR_ERR(page) =3D=3D -EBUSY) {
> -               if (!shmem_confirm_swap(mapping, index, swap))
> -                       return -EEXIST;

Do you intentionally remove checking shmem_confirm_swap()?
I am not sure I am following.

> +       if (IS_ERR_OR_NULL(page)) {
> +               if (!page)
> +                       error =3D -ENOMEM;
>                 else
> -                       return -EINVAL;
> -       } else if (!page) {
> -               error =3D -ENOMEM;
> +                       error =3D -EINVAL;

The resulting code is a bit hard to read in diff. In plan source it is like=
:

        if (IS_ERR_OR_NULL(page)) {
                if (!page)
                        error =3D -ENOMEM;
                else
                        error =3D -EINVAL;
                goto failed;
        } else {
                folio =3D page_folio(page);
                if (fault_type && result !=3D SWAP_CACHE_HIT) {
                        *fault_type |=3D VM_FAULT_MAJOR;
                        count_vm_event(PGMAJFAULT);
                        count_memcg_event_mm(fault_mm, PGMAJFAULT);
                }
        }

First of all, if the first always "goto failed", the second else is not nee=
ded.
The whole else block can be flatten one indentation.

 if (IS_ERR_OR_NULL(page)) {
                if (!page)
                        error =3D -ENOMEM;
                else
                        error =3D -EINVAL;
                goto failed;
  } else {

Can be rewrite as following with less indentation:

if (!page) {
     error =3D -ENOMEM;
     goto failed;
}
if (IS_ERR(page)) {
     error =3D -EINVAL;
     goto failed;
}
/* else block */

Am I missing something and misreading your code?

Chris