Received-SPF: pass (google.com: domain of linux-nfs-owner@vger.kernel.org designates 23.128.96.32 as permitted sender) client-ip=23.128.96.32;
MIME-Version: 1.0
References: <20231018100000.2453965-1-amir73il@gmail.com> <20231018100000.2453965-4-amir73il@gmail.com>
 <20231019144026.2qypsldg5hlca5zc@quack3>
In-Reply-To: <20231019144026.2qypsldg5hlca5zc@quack3>
From:   Amir Goldstein <amir73il@gmail.com>
Date:   Thu, 19 Oct 2023 18:22:42 +0300
Message-ID: <CAOQ4uxirRaQj8ix3aEA+SfUgXa55zaCQM3Rq+U26iqxjqnJZbA@mail.gmail.com>
Subject: Re: [PATCH 3/5] exportfs: make ->encode_fh() a mandatory method for
 NFS export
To:     Jan Kara <jack@suse.cz>
Cc:     Jeff Layton <jlayton@kernel.org>,
        Chuck Lever <chuck.lever@oracle.com>,
        Christian Brauner <brauner@kernel.org>,
        linux-fsdevel@vger.kernel.org, linux-nfs@vger.kernel.org,
        David Sterba <dsterba@suse.com>,
        Luis de Bethencourt <luisbg@kernel.org>,
        Salah Triki <salah.triki@gmail.com>,
        Gao Xiang <xiang@kernel.org>, Chao Yu <chao@kernel.org>,
        "Theodore Ts'o" <tytso@mit.edu>,
        Andreas Dilger <adilger.kernel@dilger.ca>,
        Jaegeuk Kim <jaegeuk@kernel.org>,
        OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>,
        Dave Kleikamp <shaggy@kernel.org>,
        David Woodhouse <dwmw2@infradead.org>,
        Richard Weinberger <richard@nod.at>,
        Anton Altaparmakov <anton@tuxera.com>,
        Konstantin Komarov <almaz.alexandrovich@paragon-software.com>,
        Steve French <sfrench@samba.org>,
        Phillip Lougher <phillip@squashfs.org.uk>,
        Evgeniy Dushistov <dushistov@mail.ru>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Oct 19, 2023 at 5:40=E2=80=AFPM Jan Kara <jack@suse.cz> wrote:
>
> On Wed 18-10-23 12:59:58, Amir Goldstein wrote:
> > export_operations ->encode_fh() no longer has a default implementation =
to
> > encode FILEID_INO32_GEN* file handles.
> >
> > Rename the default helper for encoding FILEID_INO32_GEN* file handles t=
o
> > generic_encode_ino32_fh() and convert the filesystems that used the
> > default implementation to use the generic helper explicitly.
> >
> > This is a step towards allowing filesystems to encode non-decodeable fi=
le
> > handles for fanotify without having to implement any export_operations.
> >
> > Signed-off-by: Amir Goldstein <amir73il@gmail.com>
>
> Just one typo cleanup. Also I agree we need a "nop" variant of
> generic_encode_ino32_fh() or move this to fs/libfs.c like e.g.
> generic_fh_to_dentry().
>

I did this:

 /*
  * Generic helpers for filesystems.
  */
+#ifdef CONFIG_EXPORTFS
+int generic_encode_ino32_fh(struct inode *inode, __u32 *fh, int *max_len,
+                           struct inode *parent);
+#else
+#define generic_encode_ino32_fh NULL
+#endif

I like it better than moving to fs/libfs.c, because if CONFIG_EXPORTFS
is not defined, no code should be calling generic_encode_ino32_fh().

It might be a good idea to define exportfs_can_*() helpers to false
when CONFIG_EXPORTFS is not defined, but at least for fanotify,
this is not relevant because fanotify selects EXPORTFS.

> > diff --git a/Documentation/filesystems/porting.rst b/Documentation/file=
systems/porting.rst
> > index 4d05b9862451..197ef78a5014 100644
> > --- a/Documentation/filesystems/porting.rst
> > +++ b/Documentation/filesystems/porting.rst
> > @@ -1045,3 +1045,12 @@ filesystem type is now moved to a later point wh=
en the devices are closed:
> >  As this is a VFS level change it has no practical consequences for fil=
esystems
> >  other than that all of them must use one of the provided kill_litter_s=
uper(),
> >  kill_anon_super(), or kill_block_super() helpers.
> > +
> > +---
> > +
> > +**mandatory**
> > +
> > +export_operations ->encode_fh() no longer has a default implementation=
 to
> > +encode FILEID_INO32_GEN* file handles.
> > +Fillesystems that used the default implementation may use the generic =
helper
>    ^^^ Filesystems
>

Thanks!
Amir.