From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Paul E. McKenney" Subject: Re: [PATCH tip/core/rcu 11/14] bonding/bond_main: Apply ACCESS_ONCE() to avoid sparse false positive Date: Sat, 16 Nov 2013 07:21:19 -0800 Message-ID: <20131116152119.GB4138@linux.vnet.ibm.com> References: <20131116003946.GA316@linux.vnet.ibm.com> <1384562417-817-1-git-send-email-paulmck@linux.vnet.ibm.com> <1384562417-817-11-git-send-email-paulmck@linux.vnet.ibm.com> <5286F550.10900@gmail.com> Reply-To: paulmck@linux.vnet.ibm.com Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Cc: Stephen Hemminger , tglx@linutronix.de, laijs@cn.fujitsu.com, edumazet@google.com, "David S. Miller" , peterz@infradead.org, fweisbec@gmail.com, bridge@lists.linux-foundation.org, linux-kernel@vger.kernel.org, rostedt@goodmis.org, josh@joshtriplett.org, dhowells@redhat.com, sbw@mit.edu, niv@us.ibm.com, netdev@vger.kernel.org, mathieu.desnoyers@efficios.com, dipankar@in.ibm.com, darren@dvhart.com, akpm@linux-foundation.org, mingo@kernel.org To: Ding Tianhong Return-path: Content-Disposition: inline In-Reply-To: <5286F550.10900@gmail.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: bridge-bounces@lists.linux-foundation.org Errors-To: bridge-bounces@lists.linux-foundation.org List-Id: netdev.vger.kernel.org On Sat, Nov 16, 2013 at 12:32:16PM +0800, Ding Tianhong wrote: > =E4=BA=8E 2013/11/16 8:40, Paul E. McKenney =E5=86=99=E9=81=93: > > From: "Paul E. McKenney" > > > > The sparse checking for rcu_assign_pointer() was recently upgraded > > to reject non-__kernel address spaces. This also rejects __rcu, > > which is almost always the right thing to do. However, the uses in > > bond_change_active_slave() and __bond_release_one() are legitimate: > > They are assigning a pointer to an element from an RCU-protected list > > (or a NULL pointer), and all elements of this list are already visibl= e > > to caller. > > > > This commit therefore silences these false positives either by launde= ring > > the pointers using ACCESS_ONCE() as suggested by Eric Dumazet and Jos= h > > Triplett, or by using RCU_INIT_POINTER() for NULL pointer assignments= . >=20 > I think it is fit for net-next. Thank you! If this is queued there, I would be happy to drop it from my tree. There are no dependencies on anything in my tree. Thanx, Paul > > Reported-by: kbuild test robot > > Signed-off-by: Paul E. McKenney > > Cc: Stephen Hemminger > > Cc: "David S. Miller" > > Cc: bridge@lists.linux-foundation.org > > Cc: netdev@vger.kernel.org > > --- > > drivers/net/bonding/bond_main.c | 5 +++-- > > 1 file changed, 3 insertions(+), 2 deletions(-) > > > > diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bo= nd_main.c > > index 72df399c4ab3..bbd7fd3e46fe 100644 > > --- a/drivers/net/bonding/bond_main.c > > +++ b/drivers/net/bonding/bond_main.c > > @@ -890,7 +890,8 @@ void bond_change_active_slave(struct bonding *bon= d, struct slave *new_active) > > if (new_active) > > bond_set_slave_active_flags(new_active); > > } else { > > - rcu_assign_pointer(bond->curr_active_slave, new_active); > > + /* Both --rcu and visible, so ACCESS_ONCE() is OK. */ > > + ACCESS_ONCE(bond->curr_active_slave) =3D new_active; > > } > > =20 > > if (bond->params.mode =3D=3D BOND_MODE_ACTIVEBACKUP) { > > @@ -1801,7 +1802,7 @@ static int __bond_release_one(struct net_device= *bond_dev, > > } > > =20 > > if (all) { > > - rcu_assign_pointer(bond->curr_active_slave, NULL); > > + RCU_INIT_POINTER(bond->curr_active_slave, NULL); > > } else if (oldcurrent =3D=3D slave) { > > /* > > * Note that we hold RTNL over this sequence, so there >=20