Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Wed, 28 Sep 2022 13:27:20 -0700 (PDT)
From:   Hugh Dickins <hughd@google.com>
To:     =?ISO-8859-15?Q?Thomas_Wei=DFschuh?= <linux@weissschuh.net>
cc:     Hugh Dickins <hughd@google.com>,
        Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org,
        =?ISO-8859-15?Q?Thomas_Wei=DFschuh?= 
        <thomas.weissschuh@amadeus.com>
Subject: Re: [PATCH] tmpfs: ensure O_LARGEFILE with generic_file_open()
In-Reply-To: <20220928104535.61186-1-linux@weissschuh.net>
Message-ID: <4d5a721-cffe-cc7-8023-ae86ba35cb7d@google.com>
References: <20220928104535.61186-1-linux@weissschuh.net>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="-1463760895-1420712399-1664396855=:21219"
Precedence: bulk

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

---1463760895-1420712399-1664396855=:21219
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: QUOTED-PRINTABLE

On Wed, 28 Sep 2022, Thomas Wei=C3=9Fschuh wrote:
> From: Thomas Wei=C3=9Fschuh <thomas.weissschuh@amadeus.com>
>=20
> Without this check open() will open large files on tmpfs although
> O_LARGEFILE was not specified. This is inconsistent with other
> filesystems.
> Also it will later result in EOVERFLOW on stat() or EFBIG on write().
>=20
> Link: https://lore.kernel.org/lkml/76bedae6-22ea-4abc-8c06-b424ceb39217@t=
-8ch.de/
> Signed-off-by: Thomas Wei=C3=9Fschuh <thomas.weissschuh@amadeus.com>

Acked-by: Hugh Dickins <hughd@google.com>

Thanks! I think you deserve some special award for finding and fixing
such an ancient bug/inconsistency - dating back to early v2.4 it seems.

But only affecting 32-bit; and since we've happily lived with it so long,
and the fix does change behaviour for userspace, better not to Cc stable.

There are some other filesystems still behaving as tmpfs was (ramfs,
hugetlbfsi, and I didn't look further); but we do make more of an effort
to keep tmpfs in line with the major filesystems - thank you.

> ---
>  mm/shmem.c | 1 +
>  1 file changed, 1 insertion(+)
>=20
> diff --git a/mm/shmem.c b/mm/shmem.c
> index 42e5888bf84d..902c5550fabc 100644
> --- a/mm/shmem.c
> +++ b/mm/shmem.c
> @@ -3876,6 +3876,7 @@ EXPORT_SYMBOL(shmem_aops);
> =20
>  static const struct file_operations shmem_file_operations =3D {
>  =09.mmap=09=09=3D shmem_mmap,
> +=09.open=09=09=3D generic_file_open,
>  =09.get_unmapped_area =3D shmem_get_unmapped_area,
>  #ifdef CONFIG_TMPFS
>  =09.llseek=09=09=3D shmem_file_llseek,
>=20
> base-commit: f76349cf41451c5c42a99f18a9163377e4b364ff
> --=20
> 2.37.3
---1463760895-1420712399-1664396855=:21219--