From: David Howells <dhowells@redhat.com>
In-Reply-To: <20070219115955.GB91@tv-sign.ru> 
References: <20070219115955.GB91@tv-sign.ru>  <20070218214359.GA4226@tv-sign.ru> <8462.1171884453@redhat.com> 
To: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Andrew Morton <akpm@linux-foundation.org>, Jarek Poplawski <jarkao2@o2.pl>,
       "David S. Miller" <davem@davemloft.net>, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/3] net/bridge/br_if.c: don't use _WORK_NAR 
Date: Mon, 19 Feb 2007 13:15:00 +0000
Message-ID: <15331.1171890900@redhat.com>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1365
Lines: 35

Oleg Nesterov <oleg@tv-sign.ru> wrote:

> > Called by what?  Something outside of br_if.c?
> 
> No. if cancel_delayed_work() fails, the work may sit pending in cwq->worklist,
> or it may be running right now, waiting for rtnl_mutex.

OIC.  I understood "called" to mean "scheduled", but that's not what you meant.

> > Hmmm... cancel_delayed_work() in del_nbp() probably ought to be followed by
> > a flush_scheduled_work().
> 
> Yes, but this deadlocks: we hold rtnl_mutex, and work->func() takes it too.

Oh, yuck!

Hmmm...  You've got a work_struct (well, a delayed_work actually) - can you
just punt the destruction of the object over to keventd to perform, I wonder?

The big problem with that that I see is that the workqueue facility has no
guards in place against a work_struct's handler function running on several
CPUs at once in response to the same work_struct.

> I think the fix should be so that port_carrier_check() does get/put on
> "struct net_bridge_port" (container), but not on "struct net_device", and

I'm not sure how this helps.  You still have to get rid of the net_device at
some point.

David
-
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/