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;
MIME-Version: 1.0
References: <CAOQ4uxh7i_s4R9pFJPENALdWGG5-dDhqPLEUXuJqSoHraktFiA@mail.gmail.com>
 <cover.1687884029.git.nabijaczleweli@nabijaczleweli.xyz> <8827a512f0974b9f261887d344c3b1ffde7b21e5.1687884031.git.nabijaczleweli@nabijaczleweli.xyz>
In-Reply-To: <8827a512f0974b9f261887d344c3b1ffde7b21e5.1687884031.git.nabijaczleweli@nabijaczleweli.xyz>
From:   Amir Goldstein <amir73il@gmail.com>
Date:   Tue, 27 Jun 2023 21:10:09 +0300
Message-ID: <CAOQ4uxj3j7gMJSojkdfe+8fQrKtJtY7wBY1UOHtQUuQ_WMjObA@mail.gmail.com>
Subject: Re: [PATCH v3 1/3] splice: always fsnotify_access(in),
 fsnotify_modify(out) on success
To:     =?UTF-8?Q?Ahelenia_Ziemia=C5=84ska?= 
        <nabijaczleweli@nabijaczleweli.xyz>
Cc:     Alexander Viro <viro@zeniv.linux.org.uk>,
        Christian Brauner <brauner@kernel.org>,
        linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
        Jan Kara <jack@suse.cz>,
        Chung-Chiang Cheng <cccheng@synology.com>, ltp@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Tue, Jun 27, 2023 at 7:55=E2=80=AFPM Ahelenia Ziemia=C5=84ska
<nabijaczleweli@nabijaczleweli.xyz> wrote:
>
> The current behaviour caused an asymmetry where some write APIs
> (write, sendfile) would notify the written-to/read-from objects,
> but splice wouldn't.
>
> This affected userspace which uses inotify, most notably coreutils
> tail -f, to monitor pipes.
> If the pipe buffer had been filled by a splice-family function:
>   * tail wouldn't know and thus wouldn't service the pipe, and
>   * all writes to the pipe would block because it's full,
> thus service was denied.
> (For the particular case of tail -f this could be worked around
>  with ---disable-inotify.)
>
> Fixes: 983652c69199 ("splice: report related fsnotify events")
> Link: https://lore.kernel.org/linux-fsdevel/jbyihkyk5dtaohdwjyivambb2gffy=
js3dodpofafnkkunxq7bu@jngkdxx65pux/t/#u
> Link: https://bugs.debian.org/1039488
> Signed-off-by: Ahelenia Ziemia=C5=84ska <nabijaczleweli@nabijaczleweli.xy=
z>
> Reviewed-by: Amir Goldstein <amir73il@gmail.com>
> ---
>  fs/splice.c | 31 ++++++++++++++-----------------
>  1 file changed, 14 insertions(+), 17 deletions(-)
>
> diff --git a/fs/splice.c b/fs/splice.c
> index 3e06611d19ae..e16f4f032d2f 100644
> --- a/fs/splice.c
> +++ b/fs/splice.c
> @@ -1154,10 +1154,8 @@ long do_splice(struct file *in, loff_t *off_in, st=
ruct file *out,
>                 if ((in->f_flags | out->f_flags) & O_NONBLOCK)
>                         flags |=3D SPLICE_F_NONBLOCK;
>
> -               return splice_pipe_to_pipe(ipipe, opipe, len, flags);
> -       }
> -
> -       if (ipipe) {
> +               ret =3D splice_pipe_to_pipe(ipipe, opipe, len, flags);
> +       } else if (ipipe) {
>                 if (off_in)
>                         return -ESPIPE;
>                 if (off_out) {
> @@ -1182,18 +1180,15 @@ long do_splice(struct file *in, loff_t *off_in, s=
truct file *out,
>                 ret =3D do_splice_from(ipipe, out, &offset, len, flags);
>                 file_end_write(out);
>
> -               if (ret > 0)
> -                       fsnotify_modify(out);
> -
>                 if (!off_out)
>                         out->f_pos =3D offset;
>                 else
>                         *off_out =3D offset;
>
> -               return ret;
> -       }
> -
> -       if (opipe) {
> +               // splice_write-> already marked out
> +               // as modified via vfs_iter_write()
> +               goto noaccessout;
> +       } else if (opipe) {
>                 if (off_out)
>                         return -ESPIPE;
>                 if (off_in) {
> @@ -1209,18 +1204,20 @@ long do_splice(struct file *in, loff_t *off_in, s=
truct file *out,
>
>                 ret =3D splice_file_to_pipe(in, opipe, &offset, len, flag=
s);
>
> -               if (ret > 0)
> -                       fsnotify_access(in);
> -
>                 if (!off_in)
>                         in->f_pos =3D offset;
>                 else
>                         *off_in =3D offset;
> +       } else
> +               return -EINVAL;
>
> -               return ret;
> -       }
> +       if (ret > 0)
> +               fsnotify_modify(out);
> +noaccessout:
> +       if (ret > 0)
> +               fsnotify_access(in);
>

As I wrote, I don't like this special case.
I prefer that we generate double IN_MODIFY than
having to maintain unreadable code.

Let's see what Jan has to say about this.

Thanks,
Amir.