From mboxrd@z Thu Jan  1 00:00:00 1970
From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Mon, 06 Apr 2020 16:34:08 +0000
Subject: Re: Upcoming: Notifications, FS notifications and fsinfo()
Message-Id: <CAHk-=wjW735UE+byK1xsM9UvpF2ubh7bCMaAOwz575U7hRCKyA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
List-Id: <keyrings.vger.kernel.org>
References: <CAJfpegs3uDzFTE4PCjZ7aZsEh8b=iy_LqO1DBJoQzkP+i4aBmw@mail.gmail.com>
 <2590640.1585757211@warthog.procyon.org.uk> <CAJfpegsXqxizOGwa045jfT6YdUpMxpXET-yJ4T8qudyQbCGkHQ@mail.gmail.com>
 <36e45eae8ad78f7b8889d9d03b8846e78d735d28.camel@themaw.net>
 <CAJfpegsCDWehsTRQ9UJYuQnghnE=M8L0_bJBTTPA+Upu87t90w@mail.gmail.com>
 <27994c53034c8f769ea063a54169317c3ee62c04.camel@themaw.net>
 <20200403111144.GB34663@gardel-login> <CAJfpeguQAw+Mgc8QBNd+h3KV8=Y-SOGT7TB_N_54wa8MCoOSzg@mail.gmail.com>
 <20200403151223.GB34800@gardel-login> <20200403203024.GB27105@fieldses.org> <20200406091701.q7ctdek2grzryiu3@ws.net.home>
In-Reply-To: <20200406091701.q7ctdek2grzryiu3@ws.net.home>
To: Karel Zak <kzak@redhat.com>
Cc: "J. Bruce Fields" <bfields@fieldses.org>, Lennart Poettering <mzxreary@0pointer.de>, Miklos Szeredi <miklos@szeredi.hu>, Ian Kent <raven@themaw.net>, David Howells <dhowells@redhat.com>, Christian Brauner <christian.brauner@ubuntu.com>, Al Viro <viro@zeniv.linux.org.uk>, dray@redhat.com, Miklos Szeredi <mszeredi@redhat.com>, Steven Whitehouse <swhiteho@redhat.com>, Jeff Layton <jlayton@redhat.com>, andres@anarazel.de, keyrings@vger.kernel.org, linux-fsdevel <linux-fsdevel@vger.kernel.org>, Linux Kernel Mailing List <linux-kernel@vger.kernel.org>, Aleksa Sarai <cyphar@cyphar.com>

On Mon, Apr 6, 2020 at 2:17 AM Karel Zak <kzak@redhat.com> wrote:
>
> On Fri, Apr 03, 2020 at 04:30:24PM -0400, J. Bruce Fields wrote:
> >
> > nfs-utils/support/misc/mountpoint.c:check_is_mountpoint() stats the file
> > and ".." and returns true if they have different st_dev or the same
> > st_ino.  Comparing mount ids sounds better.
>
> BTW, this traditional st_dev+st_ino way is not reliable for bind mounts.
> For mountpoint(1) we search the directory in /proc/self/mountinfo.

These days you should probably use openat2() with RESOLVE_NO_XDEV.

No need for any mountinfo or anything like that. Just look up the
pathname and say "don't cross mount-points", and you'll get an error
if it's a mount crossing lookup.

So this kind of thing is _not_ an argument for another kernel querying
interface.  We got a new (and better) model for a lot of this.

              Linus