From mboxrd@z Thu Jan  1 00:00:00 1970
From: Thomas Graf <tgraf@suug.ch>
Subject: Re: linux-next: net/netlink/diag.c:106 suspicious
 rcu_dereference_protected() usage!
Date: Wed, 6 Aug 2014 20:51:27 +0100
Message-ID: <20140806195127.GA26832@casper.infradead.org>
References: <CANaxB-ySsRXHPSEPOuExeKAvDX7esZ1WBZAVMTWQxtmfN2=i4w@mail.gmail.com>
 <CAHA+R7N0c83by=AAgw2nKPibYETNyv6JTEWnu5XOo-67fHMELw@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Andrey Wagin <avagin@gmail.com>, netdev <netdev@vger.kernel.org>
To: Cong Wang <cwang@twopensource.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from casper.infradead.org ([85.118.1.10]:43937 "EHLO
	casper.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1754494AbaHFTv3 (ORCPT
	<rfc822;netdev@vger.kernel.org>); Wed, 6 Aug 2014 15:51:29 -0400
Content-Disposition: inline
In-Reply-To: <CAHA+R7N0c83by=AAgw2nKPibYETNyv6JTEWnu5XOo-67fHMELw@mail.gmail.com>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On 08/06/14 at 10:52am, Cong Wang wrote:
> On Wed, Aug 6, 2014 at 10:29 AM, Andrey Wagin <avagin@gmail.com> wrote:
> > [   40.072199] ===============================
> > [   40.072350] [ INFO: suspicious RCU usage. ]
> > [   40.072501] 3.16.0-next-20140806 #1 Not tainted
> > [   40.072659] -------------------------------
> > [   40.072807] net/netlink/diag.c:106 suspicious
> > rcu_dereference_protected() usage!
> > [   40.073047]
> > other info that might help us debug this:
> >
> > [   40.073276]
> > rcu_scheduler_active = 1, debug_locks = 0
> > [   40.073494] 4 locks held by criu/2838:
> > [   40.073635]  #0:  (sock_diag_mutex){+.+.+.}, at:
> > [<ffffffff81689afb>] sock_diag_rcv+0x1b/0x40
> > [   40.074226]  #1:  (sock_diag_table_mutex){+.+.+.}, at:
> > [<ffffffff81689c6d>] sock_diag_rcv_msg+0x6d/0x140
> > [   40.074803]  #2:  (nlk->cb_mutex){+.+.+.}, at: [<ffffffff816a28f1>]
> > netlink_dump+0x21/0x2d0
> > [   40.075382]  #3:  (nl_table_lock){.+.?..}, at: [<ffffffffa0286601>]
> > netlink_diag_dump+0x31/0xb0 [netlink_diag]
> > [   40.076351]
> 
> 
> Looks like we should hold rcu_read_lock() before calling __netlink_diag_dump().

netlink_diag_dump() still acquires nl_table_lock which is pointless as
a separate mutex has been introduced to protect mutations. I will send
a patch to RCU'ify it properly.