xfs
[Top] [All Lists]

Re: xfs_file_ioctl / xfs_freeze: BUG: warning at kernel/mutex-debug.c:80

To: Andrew Morton <akpm@xxxxxxxx>
Subject: Re: xfs_file_ioctl / xfs_freeze: BUG: warning at kernel/mutex-debug.c:80/debug_mutex_unlock()
From: Ingo Molnar <mingo@xxxxxxx>
Date: Tue, 9 Jan 2007 07:41:13 +0100
Cc: David Chinner <dgc@xxxxxxx>, linux-kernel Mailing List <linux-kernel@xxxxxxxxxxxxxxx>, xfs@xxxxxxxxxxx, Arjan van de Ven <arjan@xxxxxxxxxxxxx>, Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
In-reply-to: <20070108155636.a68dce33.akpm@xxxxxxxx>
References: <20070104001420.GA32440@xxxxxxxxxxxxxxx> <20070107213734.GS44411608@xxxxxxxxxxxxxxxxx> <20070108155636.a68dce33.akpm@xxxxxxxx>
Sender: xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.4.2.2i
* Andrew Morton <akpm@xxxxxxxx> wrote:

> > Revert bd_mount_mutex back to a semaphore so that xfs_freeze -f 
> > /mnt/newtest; xfs_freeze -u /mnt/newtest works safely and doesn't 
> > produce lockdep warnings.
> 
> Sad.  The alternative would be to implement 
> mutex_unlock_dont_warn_if_a_different_task_did_it().  Ingo?  Possible?

i'd like to avoid it as much as i'd like to avoid having to add 
spin_unlock_dont_warn_if_a_different_task_did_it(). Unlocking by a 
different task is usually a sign of messy locking and bugs lurking. Is 
it really true that XFS's use of bd_mount_mutex is safe and justified?

        Ingo


<Prev in Thread] Current Thread [Next in Thread>