Subject: Re: [PATCH: 2.6.20-rc4-mm1] JFS: Avoid deadlock introduced by
	explicit I/O plugging
From: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
To: Josef Sipek <jsipek@fsl.cs.sunysb.edu>
Cc: Jens Axboe <jens.axboe@oracle.com>,
       JFS Discussion <jfs-discussion@lists.sourceforge.net>,
       fsdevel <linux-fsdevel@vger.kernel.org>,
       linux-kernel <linux-kernel@vger.kernel.org>,
       Nick Piggin <nickpiggin@yahoo.com.au>
In-Reply-To: <20070118063019.GA31164@filer.fsl.cs.sunysb.edu>
References: <1169074549.10560.10.camel@kleikamp.austin.ibm.com>
	 <20070118063019.GA31164@filer.fsl.cs.sunysb.edu>
Content-Type: text/plain
Date: Thu, 18 Jan 2007 08:15:30 -0600
Message-Id: <1169129730.7295.10.camel@kleikamp.austin.ibm.com>
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1070
Lines: 32

On Thu, 2007-01-18 at 01:30 -0500, Josef Sipek wrote:
> On Wed, Jan 17, 2007 at 04:55:49PM -0600, Dave Kleikamp wrote:

> >  /*
> >   *	jfs_lock.h
> > @@ -42,6 +43,7 @@ do {							\
> >  		if (cond)				\
> >  			break;				\
> >  		unlock_cmd;				\
> > +		blk_replug_current_nested();		\
> >  		schedule();				\
> >  		lock_cmd;				\
> >  	}						\
> 
> Is {,un}lock_cmd a macro? ...

They are the commands passed into this macro (as arguments) to
release/take a lock.  This is a home-grown wait_on_event type macro
where the condition must be checked while holding a lock.  And the
commands passed in are themselves macros.  The jfs code could probably
be cleaned up a bit as far as excessive use of macros for locking, but
it's been working for a few years with few problems.

-- 
David Kleikamp
IBM Linux Technology Center

-
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/