linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Trond Myklebust <trond.myklebust@fys.uio.no>
To: Aurelien Degremont <aurelien.degremont@cea.fr>
Cc: viro@zeniv.linux.org.uk, linux-fsdevel@vger.kernel.org,
	linux-kernel@vger.kernel.org, nfs@lists.sourceforge.net,
	Jacques-Charles Lafoucriere <jc.lafoucriere@cea.fr>
Subject: Re: NFS superblock sharing implies mount flags bug
Date: Mon, 20 Mar 2006 13:05:46 -0500	[thread overview]
Message-ID: <1142877946.7991.2.camel@lade.trondhjem.org> (raw)
In-Reply-To: <441EAC05.7020903@cea.fr>

On Mon, 2006-03-20 at 14:20 +0100, Aurelien Degremont wrote:
> Nobody is interested by this issue ?
> It could be easily reproduced. All recent versions are concerned.
> 
> Aurelien

Check the linux-fsdevel and lkml archives.

There has been plenty of work on this issue both by Herbert Poetzl and
(more recently) by Christoph Hellwig.

Cheers,
  Trond



> Aurelien Degremont wrote:
> > Hello
> > 
> > I'm facing incorrect using of mount flags when dealing with NFS mounts 
> > and I think it could be seen as a bug.
> > 
> > The error occurs when mounting the same NFS export many times, on the 
> > same machine but *with different mount flags*, particularly concerning 
> > RO/RW flags.
> > 
> > As the NFS client code re-uses superblocks when it detects that it is 
> > the same export (same server/same port/same exported directory) and that 
> > the read-only flag is managed as a per-superblock flag, if a NFS exports 
> > is mounted a second time, the superblock of the first mount is re-used 
> > and the specified mount flag is ignored.
> > 
> > # mount foo:/bar /bar_ro -o ro
> > # mount foo:/bar /bar_rw -o rw
> > $ touch /bar_rw/bar
> > touch: cannot touch `/bar_rw/bar': Read-only file system
> > 
> > Ideally, the best solution to fix this is to move the RDONLY flag from 
> > its per-superblock basis to a per-mountpoint (vfsmount) basis. I do not 
> > know is there is a something that prevent that except that this implies 
> > many changes as many codes do not use macros but access s_flags directly.
> > 
> > It seems quite clear that the superblock sharing couldn't be changed (to 
> > avoid incoherency, inode aliasing and so on...) ?
> > 
> > Do you have a (better) solution ?
> > I can help if needed.
> > 
> > 
> > Cordially
> > 
> 
> 


      reply	other threads:[~2006-03-20 18:06 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-03-15 10:37 NFS superblock sharing implies mount flags bug Aurelien Degremont
2006-03-20 13:20 ` Aurelien Degremont
2006-03-20 18:05   ` Trond Myklebust [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1142877946.7991.2.camel@lade.trondhjem.org \
    --to=trond.myklebust@fys.uio.no \
    --cc=aurelien.degremont@cea.fr \
    --cc=jc.lafoucriere@cea.fr \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=nfs@lists.sourceforge.net \
    --cc=viro@zeniv.linux.org.uk \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).