Re: DoS with POSIX file locks?

From: Miklos Szeredi <miklos@szeredi.hu>
To: bfields@fieldses.org
Cc: matthew@wil.cx, linux-fsdevel@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: DoS with POSIX file locks?
Date: Tue, 21 Mar 2006 07:38:56 +0100	[thread overview]
Message-ID: <E1FLaW8-0001hD-00@dorka.pomaz.szeredi.hu> (raw)
In-Reply-To: <20060320203556.GC31512@fieldses.org> (bfields@fieldses.org)

> > > > Things look fairly straightforward if the accounting is done in
> > > > files_struct instead of task_struct.  At least for POSIX locks.  I
> > > > haven't looked at flocks or leases yet.
> > > 
> > > I was thinking that would work, yes.  It might not be worth worrying
> > > about accounting for leases/flocks since each process can only have one
> > > of those per open file anyway.
> > 
> > Here's a minimally tested patch.  The only tricky part is when the
> > unlock splits an existing lock in two.
> > 
> > Also the limit checking is sloppy when the lock is split, and in that
> > case allows the counter to go one above the limit.
> 
> Do you need to handle blocks as well as applied locks?

I don't think so, because each task may only have at most one block at
a time.

Miklos