Date: Wed, 1 Sep 2004 18:50:43 +0200
From: Frank van Maarseveen <frankvm@xs4all.nl>
To: linux-kernel@vger.kernel.org
Subject: lookup() on non-directories.
Message-ID: <20040901165043.GA31757@janus>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.4.1i
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1318
Lines: 34

The reiserfs4 and openat() discussion made me thinking. This is different:

* /dev/hda1 -> /dev/hda/1
  => /dev/hda/1 causes a lookup in the partition table of /dev/hda

* /dev/hda/1/. is a "root inode lookup" in the superblock inside /dev/hda/1: no
  explicit mount/umount but just the usual refcounting. File system type? the
  kernel already knows about this stuff: needed for mounting the root-fs.
  /dev/hda/1 need not be a special file.

* /usr/X11R6/bin/xterm/... ?
  Kernel has some knowledge about ELF (fs/binfmt_*) so this could mean
  something.

So what's the concept?

lookup() on files are just _interpretations_ of the file based on
knowledge the kernel already has. Interpretations might require
access( , X_OK) == 0 for sanity (in addition to R_OK and/or W_OK).

This way, user mounts/automounter would no longer be a requirement for
the user in order to get access to the usual things like USB storage,
CDROM, DVD, or images of those.

yes I know, this does not address the samba or reiserfs4 issues: openat()
name space now becomes orthogonal.

-- 
Frank
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/